Cyclophosphamide induces an early wave of acrolein-independent apoptosis in the urothelium

Hemorrhagic cystitis (HC) affects a significant number of patients undergoing cyclophosphamide (CP) chemotherapy despite treatment with 2-mercaptoethane sulfonate (Mesna) to inactivate the metabolite acrolein. While the mechanism is unknown, there is clearly acrolein-independent damage to the urothelium. In this study we have explored the induction of apoptosis in the urothelium as a marker of damage and the mechanism underlying the acrolein-independent apoptosis. Two waves of apoptosis (measured as caspase-3/7 activity and Poly (ADP-ribosyl) polymerase (PARP) cleavage) were detected following CP administration, one peaking at 2 h and a second at 48 h. The first wave was not blocked by Mesna, indicating it was independent of acrolein. Caspase-1 was also active at 2 h and activation of caspase-3/7 was blocked by a caspase-1 inhibitor, but not an IL-1 receptor antagonist, suggesting the direct activation of caspase-3/7 by caspase-1 without the need for IL-1 as an intermediate. Our results indicate that CP initiates an early, acrolein-independent wave of apoptosis that results from direct cleavage of caspase-3/7 by caspase-1.


INTRODUCTION
Cyclophosphamide (CP) is a widely used chemotherapeutic agent that causes acute hemorrhagic cystitis (HC) [1].A metabolite of CP, acrolein, is largely responsible, leading oncologists to concomitantly administer 2-mer-captoethane sulfonate (Mesna), an orally-available drug that masks acrolein's toxic effect.The use of Mesna has dramatically reduced, but not eliminated, HC [1][2][3][4][5][6][7] following CP chemotherapy.The mechanism for HC in the presence of Mesna has not been defined but clearly involves damage to the urothelium.A typical cellular response to damage is to initiate apoptosis in order to clear the damaged cell from the population.In this study, we have examined CP-induced damage to the urothelium in vivo using two independent markers of apoptosis: the activation of caspase(s)-3/7 and the cleavage of Poly (ADP-ribose) polymerase (PARP).Caspase-3/7 is considered an effector caspase and is activated in most models of apoptosis where it is directly involved in dismantling the cell by cleavage of structural and functional proteins.PARP is a DNA repair enzyme that is cleaved and inactivated ostensibly to conserve energy by not repairing DNA intentionally destroyed during the death process.
Recent studies have shown that damage to a cell or tissues is recognized by pattern recognition receptors which result in the activation of the inflammatory caspase, caspase-1 [8,9].We have further examined the involvement of caspase-1 in the response to CP and the mechanism of its involvement.

Animals
Adult female randomly cycling Sprague Dawley rats (175 -199 g; Harlan, Prattville AL, USA) were used in all studies.All experimental protocols were approved by the Institutional Animal Care and Use Committee at our institution.

Caspase Assays
Urothelium were obtained by scraping, similar to what has been reported elsewhere [13,14].Briefly, the isolated bladder was cut open from sphincter to dome.The dome ends were pinned and the sphincter held by forceps with a bent tip.The urothelium was gently scraped from the detrusor using a size 11 scalpel blade, twisting the blade to lift the cells.The cells were placed in ice-cold PBS, pelleted (800 × g, 10 min), resuspended in 25 µl of 10 mM MgCl 2 , 0.25% Igepal CA-630, mixed with an equal volume of 40 mM Hepes (pH 7.4), 20 mM NaCl, 2 mM EDTA, 20% Glycerol and stored at −20˚C until analyzed (<1 week).

Statistics
To test for differences over time (Figure 2), we modeled the log of percent difference from baseline (time = 0) using linear regression with fixed effects for hours from baseline, treated as categorical variables.Regression model assumptions were evaluated using residual plots and QQ-plots.p-values represent tests that percent changes from baseline were different from 0, and p values < 0.05 were considered significant.Other comparisons were conducted using non-parametric Wilcoxon rank sum tests or Student's t-test, as appropriate, and were also considered significant if p < 0.05.

There Are Two Waves of Apoptosis Following CP Administration
Caspase-3/7 activity in urothelium was followed for 72 h after the injection of 80 mg/kg CP.As shown in Figure 2(a), we detected a biphasic induction of apoptosis with an initial peak at 2 h, followed by a return to baseline at 8 h.At 16 h, levels had increased again and continued to go up until at least the 48 h time point.While we have previously shown that this dosing paradigm induces changes in physiological bladder function indicative of HC at 24 h [10], the peak at 2 h is novel and is not associated with gross morphological change (data not shown).
To confirm the 2 h result, we have examined expression of an independent apoptotic marker, cleaved PARP [16].
As shown in Figures 2(b) and (c), there was a 21.2% increase in the percentage of the urothelium cells with cleaved PARP at 2 h (% of cells under the M1 gate).

The First Wave of Caspase-3/7 Activity Is Independent of Acrolein
To access the contribution of acrolein to the two waves of apoptosis, rats were pretreated with Mesna 0.5 h prior to CP administration.Due to the short half-life of Mesna [17], rats assessed at 24 h also received additional doses of Mesna at 4 and 8 h [11,12] (Figure 1(a)).As shown in Figure 1(b), Mesna significantly (p < 0.05) reduced the caspase-3/7 activity in urothelium at 24 h but only slightly and non-significantly reduced the activity at 2 h, indicating that the early peak of apoptosis is mostly acrolein-independent while the latter wave is largely dependent on this metabolite.

A Caspase-1 Inhibitor, But Not an IL-1 Receptor Antagonist, Blocks Activation of Caspase-3/7 at 2 h
Damage to a cell or tissue often results in the activation of caspase-1.As shown in Figure 3(a), we detect significant activation of caspase-1 two hours after CP administration, further indicating urothelium damage at this time point.Working upstream, caspase-1 can cleave caspase-3/7 directly [18,19] or stimulate apoptosis through production of IL-1β [20,21].Direct luminal instillation of the caspase-1 inhibitor YVAD-AOM blocked the increase in caspase-3/7 activity stimulated by CP after 2 h (Figure 3(b)).Caspase-3/7 activity was significantly (p < 0.05) stimulated by CP in rats that were not pretreated with anakinra (an IL-1 receptor antagonist) and in ones that were pretreated (Figure 3(c)).In addition, there was no significant (p = 0.28) difference noted between the anakinra pretreated and saline rats which did not receive CP, and there was no significant (p = 0.10) difference noted between the anakinra pretreated and saline rats which did receive CP.In other words, pretreatment of rats with the IL-1β receptor antagonist anakinra had no effect (Figure 3(c)).

DISCUSSION
CP-induced damage to the urothelium is clearly involved in the induction of HC.While the metabolite acrolein is largely responsible for this, the persistence of significant levels of this malady in the presence of Mesna suggests additional damage must occur.Since damage to cells often activates apoptosis, we examined the increase of this death process as an indicator of damage in response to CP.In these studies we have used a smaller dose of CP than is often used with rodents [22,23] but one we have previously documented to induce changes in physiological bladder function indicative of HC by 24 h [10].Moreover, this dose was chosen from our institution's protocols for treating patients with acute lymphocytic leukemia.
Using caspase-3/7 activity as a measurement of apoptosis, we detected a biphasic response with waves peaking at 2 and 48 h.The peak at 2 hours is novel and the inclusion of Mesna had an insignificant effect at this time.The rise in cell death at 24 h is consistent with Jezernik et al. [24], although they did not find significant apoptosis at 1 or 6 h following CP.Given the transient nature of the early wave at 2 h it is possible that Jezernik et al. bracketed the peak and thus did not detect significant levels of cell death.Importantly, we have confirmed the 2 h result using an independent apoptotic marker; cleaved PARP, which also provided an estimate of the population effected (21.2% ± 4.9%).Consequently, our results indicate acrolein-independent damage to the uro-thelium early following CP administration and we sought to explore this change further.
Recent studies have begun to delineate how damage is sensed in a tissue and the results show a central role for pattern recognition receptors and the activation of caspase-1 [8,9].Indeed, we detected caspase-1 activated at 2 h following CP treatment, providing further evidence of tissue damage at this time.This result also suggested two possible links between activation of this inflammatory caspase and the downstream effector caspase; caspase-3/7.Caspase-1 can directly cleave caspase-3/7 [18,19] or activate IL-1β, which itself may induce apoptosis and result in the activation of caspase-3/7 [20,21].To differentiate between these possibilities we used a caspase-3/7 inhibitor (YVAD-AOM) and a IL-1 receptor antagonist (anakinra).While the caspase-1 inhibitor blocks the activation of caspase-3/7, anakinra had no effect which argues strongly for the direct activation of caspase-3/7 without IL-1β as an intermediate.While IL-1β is most certainly produced in our model, it is not implicated in the initial wave of apoptosis at 2 h.
Our results demonstrate that a single dose of CP at clinically relevant levels [10] can initiate an early wave of damage in the urothelium resulting in the activation of caspase-1 and apoptosis.How these events could precipitate later HC is unknown.Interestingly, recent work with herpes virus demonstrates that apoptosis can reactivate this virus [25,26].Considering that reactivation of BK virus (latently present in 90% of adults) has been strongly implicated in the pathology underlying HC in Mesna-treated patients [3,27,28], it is tempting to speculate that apoptotic-induced viral reactivation may be involved.
Our results indicate that CP initiates an early, acroleinindependent wave of apoptosis that results from direct cleavage of caspase-3/7 by caspase-1.

Figure 3 .
Figure 3. (a) Caspase-1 activity in rat urothelium 2 h after injection of 80 mg/kg CP or saline control.Data points are the mean ± SEM.Asterisks indicate values significantly different from the mean value at 0 h, n = 6, p < 0.05.(b) (upper) diagram of the protocol for treatment with YVAD-AOM, CP and DMSO.(b) (lower) caspase-3/7 activity in the various treatment groups.Data points are mean ± SEM.Asterisks indicate values significantly different from DMSO control; n = 6, p < 0.05.No other differences between groups were significant.(c) (upper) diagram of the protocol for treatment with Anakinra (Ana), CP and saline.(c) (lower) caspase-3/7 activity in the various treatment groups.Data points are mean ± SEM.Asterisks indicate values significantly different from saline control; n = 5 -6, p < 0.05.No other differences between groups were significant.