Interruption of Homologous Desensitization in Cyclic Guanosine 3 V , 5 V-Monophosphate Signaling Restores Colon Cancer Cytostasis by Bacterial Enterotoxins

Bacterial diarrheagenic heat-stable enterotoxins induce colon cancer cell cytostasis by targeting guanylyl cyclase C (GCC) signaling. Anticancer actions of these toxins are mediated by cyclic guanosine 3V,5V-monophosphate (cGMP)–dependent influx of Ca through cyclic nucleotide-gated channels. However, prolonged stimulation of GCC produces resistance in tumor cells to heat-stable enterotoxin–induced cytostasis. Resistance reflects rapid (tachyphylaxis) and slow (bradyphylaxis) mechanisms of desensitization induced by cGMP. Tachyphylaxis is mediated by cGMP-dependent protein kinase, which limits the conductance of cyclic nucleotide-gated channels, reducing the influx of Ca propagating the antiproliferative signal from the membrane to the nucleus. In contrast, bradyphylaxis is mediated by cGMP-dependent allosteric activation of phosphodiesterase 5, which shapes the amplitude and duration of heat-stable enterotoxin–dependent cyclic nucleotide accumulation required for cytostasis. Importantly, interruption of tachyphylaxis and bradyphylaxis restores cancer cell cytostasis induced by heat-stable enterotoxins. Thus, regimens that incorporate cytostatic bacterial enterotoxins and inhibitors of cGMP-mediated desensitization offer a previously unrecognized therapeutic paradigm for treatment and prevention of colorectal cancer. (Cancer Res 2005; 65(23): 11129-35)


Introduction
Colorectal cancer is the third most common, and the second most deadly, cancer in the developed world (1,2).The mortality rate for colon cancer, 50%, reflects metastatic disease progression (1) and the lack of efficacious adjuvant chemotherapy (3).Indeed, f20% of patients have unresectable disease at presentation and the majority of patients (>90%) that develop metastases (f33%) do not benefit from current pharmacotherapeutic interventions (1,3).Major obstacles to the development of effective therapeutic regimens include the genetic and phenotypic heterogeneity of colorectal tumors (4) and the emergence of drug-induced adaptive escape mechanisms in cancer cells (5).Thus, individualized therapy, more effective molecular targets, and strategies to circumvent drug resistance are paramount for future therapeutic paradigms for colon cancer (6,7).
Bacterial heat-stable enterotoxins induce secretory diarrhea in endemic populations, travelers, and animal herds (8) by serving as superagonists for the intestine-specific receptor guanylyl cyclase C (GCC; ref. 9).Heat-stable enterotoxins, which have evolved to facilitate bacterial dissemination and propagation, exemplify molecular mimicry of the endogenous hormones guanylin and uroguanylin, which mediate autocrine/paracrine control of intestinal fluid and electrolyte homeostasis by activating GCC and inducing cyclic guanosine 3V ,5V -monophosphate (cGMP)-dependent chloride efflux through the cystic fibrosis regulator channel (10)(11)(12).Intriguingly, longitudinal exposure to heat-stable enterotoxin-producing bacteria seems to protect endemic populations against colon cancer by reducing rates of enterocyte proliferation (13) and intestinal tumorigenesis (14).In contrast to fluid and electrolyte secretion, regulation of intestinal cell proliferation by heat-stable enterotoxin-induced activation of GCC consists of cGMP-dependent stimulation of Ca 2+ entry through cyclic nucleotide-gated (CNG) channels (13).Heat-stable enterotoxin-induced Ca 2+ currents, in turn, are coupled to suppression of DNA synthesis (13) and tumor cell cytostasis (15).
Normally, GCC is selectively expressed in apical membranes of enterocytes, ''outside'' mucosal cell tight junctions, and inaccessible to the systemic vascular compartment (10,16,17).Intestinal epithelial cells that have undergone neoplastic transformation overexpress functionally competent GCC (18) displayed on their surface during metastatic dissemination to extraintestinal tissues, making it paradoxically accessible to the systemic vascular compartment (19,20).Indeed, GCC represents a unique target for selectively delivering imaging and therapeutic agents to metastatic colorectal tumors in vivo (16,17).Moreover, GCC agonists have been proposed as novel cytostatic agents for targeted therapy for colorectal cancer metastases (13).However, receptor desensitization (21)(22)(23) and activation of phosphodiesterases (21,24) represent mechanisms by which colorectal cancer cells could develop resistance to cGMP-dependent cytostasis, limiting the therapeutic potential of GCC ligands.
The present study reveals the previously unrecognized emergence of homologous desensitization of GCC-mediated cell cycle regulation in human colon cancer cells.Thus, elevations in intracellular cGMP ([cGMP i ]) induce rapid (tachyphylaxis) and slow (bradyphylaxis) mechanisms of desensitization, imposed by the integrated regulation of discreet cGMP-dependent effectors, which prevent GCC-mediated cytostasis.Importantly, interruption of the molecular mechanisms underlying tachyphylaxis and bradyphylaxis permits sustained inhibition of cancer cell proliferation by GCC ligands without the development of escape or resistance.
Cell proliferation.Proliferation of cancer cells was quantified in 96-well plates by [methyl-3 H]thymidine (0.2 ACi/well) incorporation into DNA (15).Cells were pulse-labeled (3 hours) with [ 3 H]thymidine at the end of respective stimulation periods.For studies examining cancer cell cytostasis, f50,000 cells per well were plated, permitted to recover for 6 hours, and synchronized for 18 hours with FBS-free DMEM.Then, cells were stimulated to proliferate (up to 34 hours) by adding 10% FBS, in the presence of the indicated treatments.For studies examining acute regulation of DNA synthesis, exponentially growing cancer cells (f60% confluent) were synchronized by FBS starvation in EMEM for 48 hours followed by stimulation of DNA synthesis with 10 mmol/L L-glutamine (in EMEM, for f24 hours).Where appropriate, pretreatments with heat-stable enterotoxins and/or other agents were done during the stimulation period and terminated by washing cells with EMEM (37jC, thrice) at the 20th hour following glutamine addition.Then, treatments in L-glutamine-containing EMEM were added to cells f15 minutes before [methyl-3 H]thymidine for quantification of effects on DNA synthesis (13,15).Unless otherwise indicated, effects on DNA synthesis were expressed as the percentage of parallel control cultures employing the corresponding vehicle.
Flow cytometry.T84 cells (f10 6 per well) were plated in six-well plates, permitted to recover for 6 hours, and synchronized for 18 hours by serum starvation.Then, cell proliferation was induced with 10% FBS for 24 hours, in the presence of the indicated treatments.Following incubations, cells were collected by trypsinization, washed with PBS (thrice), fixed with icecold ethanol (75%), and stained with a propidium iodide-containing solution (50 Ag/mL propidium iodide, 100 Ag/mL RNase A, 1 mmol/L EDTA, and 0.1% Triton X-100).Quantification of the percentage of cells in each phase of the cell cycle was done using a Coulter EPICS XL-MCL flow cytometer (15).
Cyclic guanosine 3V ,5V -monophosphate.Studies were done in exponentially growing T84 cells (f60% confluent), and cGMP was determined by RIA (19).Cells were synchronized, stimulated with 10 mmol/L L-glutamine in EMEM, and pretreated where appropriate as described for studies examining DNA synthesis.Then, for total cellular cGMP accumulation, cells in 96-well plates were washed (thrice) and incubated for 15 minutes in EMEM at 37jC, in the presence of the indicated treatment.Reactions were terminated with ice-cold 100% ethanol, and supernatants separated from pellets by centrifugation and processed for cGMP determinations.In contrast, for guanylyl cyclase activity, cells in T-75 flasks were collected by scraping into 1 mL of TEED buffer [50 mmol/L Tris-HCl (pH 7.5), 1 mmol/L EGTA, 1 mmol/L EDTA, 1 mmol/L DTT, and 1 mmol/L phenylmethylsulfonyl fluoride], homogenized on ice by aspiration through a series of narrow gauge (20-25 gauge) needles, and centrifuged at 1,000 Â g (4jC) for 5 minutes.Supernatants were further centrifuged at 100,000 Â g (4jC) for 60 minutes to produce a pellet, which was resuspended in TEED at 1 mg protein/mL.Membrane proteins (20 Ag) were incubated for 5 minutes (37jC) in 100 AL of 50 mmol/L Tris-HCl (pH 7.5), 1 mmol/L isobutylmethylxanthine, 15 mmol/L creatine phosphate, 2.7 units of creatine phosphokinase, 4 mmol/L MgCl 2 , 1 mmol/L GTP, in the presence of the indicated treatment.Enzyme reactions were terminated by addition of 50 mmol/L sodium acetate (pH 4.0) followed by boiling for 3 minutes.cGMP production was linear with respect to time and protein concentration.
Statistics.Reverse transcription-PCR experiments were done in duplicate and repeated twice.Studies of calcium transport were done in duplicate and expressed as mean F SE of four separate experiments.All other determinations were done in triplicate and repeated at least thrice.Data are expressed as the mean F SE of a representative experiment done in triplicate.Statistical analysis was done employing the unpaired two-tailed Student's t test, and significance was assumed for P < 0.05.

Results
Human colon carcinoma cells develop resistance to heatstable enterotoxin-induced cytostasis.Heat-stable enterotoxins delayed cell cycle progression f40% (quantified by double reciprocal analysis) in human colon cancer cells (Fig. 1A), in the absence of checkpoint arrest or apoptosis (Fig. 1B; ref. 15).Cell cycle delay reflected an acute effect that underwent desensitization, because suppression of DNA synthesis was lost following chronic exposure to heat-stable enterotoxins (Fig. 1C).Desensitization of the antiproliferative effects of heat-stable enterotoxins was associated with a reduced ability of that enterotoxin to elevate [cGMP] i (Fig. 1D), which mediates cytostasis produced by GCC (13,15).Moreover, pharmacologic inhibition of phosphodiesterase 5 (PDE5), but not PDE3, potentiated heat-stable enterotoxin effects on proliferation by further elevating [cGMP] i (Fig. 1D), and restored maximal heat-stable enterotoxin-induced [cGMP] i accumulation in desensitized cells (Fig. 1D, arrows).However, suppression of DNA synthesis by heat-stable enterotoxins was only partially rescued by inhibiting PDE5 in desensitized tumor cells (Fig. 1D).Thus, colon cancer cell resistance to the antiproliferative effects of heat-stable enterotoxins reflects a complex mechanism mediated, in part, by PDE5, which shapes GCC signaling by limiting the amplitude of [cGMP] i accumulation.
Enterotoxin-induced resistance to cytostasis is mediated by tachyphylaxis and bradyphylaxis.Colon cancer cells developed resistance to the antiproliferative effects of heat-stable enterotoxins as early as 30 minutes following exposure to that ligand (Fig. 2, left).Following complete desensitization, heat-stable enterotoxins induced a paradoxical stimulation of DNA synthesis, which peaked after 3 hours of exposure to the enterotoxin (Fig. 2, left).Moreover, mechanisms of cancer cell adaptation were completely reversible within 9 hours following withdrawal of heat-stable enterotoxins (Fig. 2, left).These temporal kinetics of desensitization and recovery of the antiproliferative effects of heatstable enterotoxins are nearly identical to those of induction and decline of PDE5 activity mediated by GCC and cGMP signaling (21,25).Of importance, linear regression analysis of the onset of cellular adaptation revealed two principle components characterized by rapid (tachyphylaxis) and slow (bradyphylaxis) rates of desensitization (Fig. 2, right).
Bradyphylaxis to heat-stable enterotoxin-induced cytostasis is mediated by PDE5.Bradyphylaxis was characterized by a paradoxical stimulation of DNA synthesis (Fig. 2) associated with a similarly paradoxical decrease in [cGMP] i accumulation induced by heat-stable enterotoxins (Fig. 3A).Inhibition of PDE5, but not PDE3, prevented paradoxical stimulation of DNA synthesis (Fig. 3B) by restoring maximal heat-stable enterotoxin-induced [cGMP] i accumulation (Fig. 1D, arrows).In that context, cGMP induces durable allosteric activation of PDE5 (26,27), which could mediate paradoxical enterotoxin-dependent stimulation of DNA synthesis (Fig. 3B) and decreases in [cGMP] i accumulation (Fig. 3A) following chronic stimulation of GCC by heat-stable enterotoxins.Indeed, a membrane-permeant phosphodiesterase-resistant analogue of cGMP, 8-br-cGMP (28), which mimicked heat-stable enterotoxin-induced cytostasis and desensitization in colon cancer cells (Fig. 3C), failed to induce paradoxical stimulation of DNA synthesis in heat-stable enterotoxin-or 8-br-cGMP-desensitized cells (Fig. 3C, compare with the respective PBS condition).In addition, 8-br-cGMP was equiefficacious with the combination of heat-stable enterotoxins and a PDE5 inhibitor in partially reversing heat-stable enterotoxin-induced desensitization (Fig. 3C).Importantly, cancer cells desensitized with 8-br-cGMP, a poor allosteric activator of PDE5 (29), did not exhibit paradoxical stimulation of DNA synthesis upon heat-stable enterotoxin exposure and were insensitive to the combination of heat-stable enterotoxin and a PDE5 inhibitor (Fig. 3C).It is noteworthy that, although heat-stable enterotoxins failed to maximally stimulate GCC activity over long durations (Fig. 4A), receptor down-regulation (30) and/or decreased cGMP synthetic rates (21) did not substantially contribute to heat-stable enterotoxin resistance in intact cells, because inhibition of PDE5 activity fully prevented bradyphylaxis (Fig. 3B).Thus, PDE5 largely mediates bradyphylaxis to heat-stable enterotoxin-dependent cytostasis in colon cancer cells by upregulation of its catalytic activity induced by chronic cGMP stimulation (25)(26)(27).
Tachyphylaxis to heat-stable enterotoxin-induced cytostasis is mediated by PKG.Tachyphylaxis contributed f50% of the desensitization to heat-stable enterotoxin-induced cytostasis (y intercepts of extrapolated fitted curves, Fig. 2, right).Rapid desensitization also was mediated by cGMP (Fig. 3C) but reflected a mechanism other than synthesis or degradation of cGMP by GCC (Fig. 4A) or PDE5 (Fig. 1D), respectively.Thus, tachyphylaxis was induced by the phosphodiesterase-resistant 8-br-cGMP (28), which increased [cGMP] i independently of GCC without inducing bradyphylaxis by PDE5 (Fig. 3C).Ca 2+ entry through CNG channels expressed in colorectal cancer cells (ref.31; Fig. 4B, top) mediated heat-stable enterotoxin-induced cytostasis (Fig. 4B, bottom; ref. 13).Although [cGMP] i remained stable, this Ca 2+ current was attenuated over the time course of tachyphylaxis (Fig. 4C) and could be fully restored by inhibiting PKG but not PDE5 (Fig. 4D).Diminished Ca 2+ entry through CNG channels mediated tachyphylaxis, because protection of the cytostatic Ca 2+ current by inhibiting PKG, but not PDE5, prevented acute desensitization to the antiproliferative effects of heat-stable enterotoxins (Fig. 4D).Importantly, tachyphylaxis and the associated attenuation of Ca 2+ entry were not influenced by removal of extracellular Ca 2+ (Fig. 4D), indicating that PKG-induced desensitization is not dependent on mechanisms downstream of the tertiary messenger Figure 1.A, synchronized T84 cells were stimulated to proliferate with 10% FBS for the indicated times in the presence of vehicle (PBS, 5) or heat-stable enterotoxins (ST , 1 Amol/L, inducing maximal inhibition of cancer cell proliferation; ref. 15; E).Cells were pulse-labeled (3 hours) with [methyl- 3 H]thymidine at the end of incubations.Percentage of the maximal thymidine incorporation induced by serum (PBS, 24 hours).B, cell cycle distribution by flow cytometry (see Materials and Methods) of T84 cells treated for 24 hours with PBS or heat-stable enterotoxins (1 Amol/L).C, DNA synthesis of synchronized T84 cells was stimulated for 20 hours by L-glutamine in the presence of PBS or heat-stable enterotoxins (1 Amol/L).Cells were then washed and exposed to heat-stable enterotoxins (1 Amol/L) in L-glutamine-containing medium.After 15 minutes, [methyl -3 H]thymidine was added and incubations were continued for an additional 3 hours.D, PBS or heat-stable enterotoxins (1 Amol/L) pretreatments and inhibition of DNA synthesis by 1 Amol/L heat-stable enterotoxins was assessed as in C .In parallel incubations, total cellular cGMP accumulation was determined by stimulating cells for 15 minutes with heat-stable enterotoxins (1 Amol/L) following pretreatments.Inhibitors of cGMP-regulated, cGMP-specific PDE5 [zaprinast (ZAP ), 10 Amol/L], and cGMP-regulated, cAMP-specific PDE3 [milrinone (MLR ), 10 Amol/L] were added 15 minutes before the final application of heat-stable enterotoxins to completely inhibit the respective targeted phosphodiesterase (25,49).
Figure 2. Heat-stable enterotoxins (ST, 1 Amol/L) effects on DNA synthesis were determined in T84 cells following pretreatment (1 Amol/L heat-stable enterotoxins) and washing as indicated (left ).[ 3 H]Thymidine incorporation into DNA was done and measured as described in Materials and Methods.Linear regression analyses were generated to fit acute (0-1 hours) and chronic (1-3 hours) onset of desensitization (right ).Rates of desensitization (K ds ) were calculated from the slope of respective lines.
(Ca 2+ ) mediating cytostasis by heat-stable enterotoxins (13).Hence, in colon cancer cells, tachyphylaxis to heat-stable enterotoxininduced cytostasis is a negative feedback mechanism mediated by PKG and targeting CNG channels.
Inhibition of tachyphylaxis and bradyphylaxis restores heat-stable enterotoxin-induced cytostasis in human colon carcinoma cells.Cancer cell resistance to cytostasis induced by heat-stable enterotoxins or 8-br-cGMP was prevented by the combination of two selective pharmacologic inhibitors of PDE5 and PKG, respectively (Fig. 5A).Prevention of enterotoxin-induced desensitization, mediated by tachyphylaxis and bradyphylaxis, by these agents was additive, reflecting the contribution of both PKG and PDE5 to these mechanisms (Fig. 5A, left).In contrast, the PDE5 inhibitor zaprinast had little effect on, whereas the PKG inhibitor RP8pCPT-cGMPs completely reversed, 8-br-cGMP-induced desensitization, reflecting the dominant contribution of tachyphylaxis to this mechanism (Fig. 5A, right).Moreover, combining these inhibitors with heat-stable enterotoxins unmasked durable suppression of tumor cell proliferation (Fig. 5B).These observations show that the cell cycle delay induced by heat-stable enterotoxins (Fig. 1A) reflects the integration of cancer cell adaptation and escape superimposed on cytostasis rather than enduring cell cycle inhibition.Finally, the combination of heat-stable enterotoxins and inhibitors of PDE5 and PKG induced cytostasis but not cell cycle arrest or apoptosis (Fig. 5C), underscoring their specific action in preventing desensitization to the antiproliferative effects of GCC and cGMP (Fig. 1B; ref. 15).

Discussion
Current pharmacotherapy for colorectal cancer has limited clinical effect (3), and f50% of patients die within 5 years from diagnosis as a result of metastatic disease (1).Adjuvant chemotherapy, the best available intervention for colon cancer metastasis, increases median survival only f14 months (32).One principle mechanism underlying the minimal efficacy of current therapeutic regimens is the acquisition of tumor cell resistance to anticancer drugs (5).For example, the drug of choice in colon cancer, 5-fluorouracil, is ineffective in tumors expressing high levels of dihydropyrimidine dehydrogenase (33) or thymidylate synthase (34), which increases drug catabolism and decreases drug-induced inhibition of DNA synthesis, respectively.Moreover, tumors frequently adopt opportunistic and nonspecific but highly effective escape mechanisms against anticancer drugs, such as multidrug resistance associated with increased drug efflux following induction of ATP-dependent transporters (35).The varied molecular mechanisms underlying drug resistance underscore the importance of designing strategies to circumvent tumor cell adaptation, which affects even the most recently introduced therapeutically efficacious agents (5,7).
Cytostasis induced by bacterial diarrheagenic heat-stable enterotoxins represents a novel therapeutic approach against colon Figure 3. A, DNA synthesis and total cGMP accumulation was performed as in Fig. 1D .After pretreatments (20 hours), T84 cultures were exposed to PBS or 1 Amol/L heat-stable enterotoxins (ST ).DNA synthesis is expressed as the percentage of the PBS control of cells pretreated with PBS.B, inhibition of DNA synthesis by 1 Amol/L heat-stable enterotoxins was assessed following pretreatments with heat-stable enterotoxins (1 Amol/L) for the indicated times.Zaprinast (ZAP, 10 Amol/L) and milrinone (MLR, 10 Amol/L) were added 15 minutes before the final application of heat-stable enterotoxins.Experiments were done as described in Materials and Methods.Percentage of the PBS control of cells pretreated with 1 Amol/L heat-stable enterotoxins for 20 hours.*, P < 0.05; **, P < 0.01, compared with the PBS control.C, following pretreatment for 20 hours with PBS, 1 Amol/L heat-stable enterotoxins or 5 mmol/L 8-br-cGMP, T84 cells were reexposed to the same reagents 15 minutes before being pulse labeled (3 hours) with [methyl -3 H]thymidine to assess DNA synthesis.Where indicated, zaprinast (10 Amol/L) was added 15 minutes before the final application of heat-stable enterotoxins.
Figure 4. A, guanylyl cyclase activity was quantified as described in Materials and Methods.Membrane preparations were obtained from T84 cells following incubations with PBS or 1 Amol/L heat-stable enterotoxins (ST ).*, P < 0.05, compared with the effect of heat-stable enterotoxins in membranes from cells not preincubated with heat-stable enterotoxins.B, RT-PCR for CNG channel (top ), DNA synthesis (bottom left ), and 45 Ca 2+ influx (bottom right ) were examined as described in Materials and Methods.NTC, nontemplate control.Heat-stable enterotoxins (1 Amol/L) and the specific inhibitor of CNG channels L-cis -diltiazem (L-DLT , 200 Amol/L) were employed to examine heat-stable enterotoxin-mediated regulation of antiproliferative 45 Ca 2+ entry through CNG channels (13).Measurements of 45 Ca 2+ influx are expressed as % increase over respective controls.Acute desensitization was studied in T84 cells following preincubations with 1 Amol/L heat-stable enterotoxins (for 0-60 minutes in C ) or 1 Amol/L heat-stable enterotoxins in the presence of the indicated reagents (for 60 minutes in D ).DNA synthesis, 45 Ca 2+ influx, and cGMP accumulation induced by heat-stable enterotoxins (1 Amol/L) were examined in parallel incubations and measured as described in Materials and Methods.cGMP is expressed as the relative fold of respective PBS controls (baselines).Ca 2+ -free, Ca 2+ -deficient cell culture media (S-MEM); ZAP, the PDE5 inhibitor zaprinast (10 Amol/L); RP8pCPT, the PKG inhibitor RP8pCPT-cGMPs, employed at 100 Amol/L to completely inhibit PKG (50).
cancer metastases (13), which uniformly overexpress GCC at their surfaces (18).Heat-stable enterotoxins inhibit tumor cell proliferation by targeting the intestinal receptor GCC and inducing cGMP-dependent Ca 2+ influx through CNG channels (13,15).Of significance, expression of the endogenous ligands of GCC, guanylin and uroguanylin, is invariably lost early during neoplastic transformation in intestine (36)(37)(38), suggesting a tumor suppressor role for the pathway targeted by heat-stable enterotoxins (13).One limitation for GCC-targeted therapy is the development of negative feedback mechanisms in cGMP signaling by colon cancer cells, including (a) reduced GCC sensitivity to the ligand ( 21), (b) increased PDE5-mediated cGMP degradation (25), (c) inhibition of GCC catalytic activity by the SH3 domain of Src tyrosine kinase (22), and (d) increased internalization of hyperglycosylated GCC from the cell surface (23).
Here, homologous desensitization of heat-stable enterotoxininduced cytostasis was mediated specifically by PDE5 and PKG in colon cancer cells.PDE5 is a cGMP-regulated, cGMP-specific phosphodiesterase, whose catalytic activity is stimulated by cGMP binding to allosteric sites at the NH 2 -terminal domains (27), and by PKG-mediated phosphorylation at Ser 92 (26,39).PDE5 activation by cGMP mediated bradyphylaxis to heat-stable enterotoxins associated with paradoxical stimulation of DNA synthesis.Indeed, cGMPstimulated PDE5 limited increases in [cGMP] i required for enterotoxin-induced cytostasis in tumor cells exposed to heatstable enterotoxins over long durations.In agreement with the present findings, cGMP-dependent PDE5 activation reduced maximum stimulation of GCC by heat-stable enterotoxins in human colon cancer cells (21,25).Similarly, enhanced activation of PDE5-dependent cGMP hydrolysis was observed in the heart (40) and kidney (41) upon persistent stimulation of cGMP synthesis by natriuretic peptide receptors, particulate guanylyl cyclases homologous to GCC.Moreover, cGMP-stimulated PDE5 opposed chronic nitric oxide stimulation of soluble guanylyl cyclase in platelets and vascular smooth muscle (42,43).Together, these observations support the suggestion that desensitization, in the form of bradyphylaxis, shapes the amplitude and duration of cGMP responses in all cells coexpressing guanylyl cyclases and PDE5 (27,43).
In contrast to PDE5 and bradyphylaxis, PKG-mediated tachyphylaxis represents a previously unrecognized mechanism of cellular adaptation to GCC and cGMP signaling.PKG regulated rapid desensitization to heat-stable enterotoxin-induced cytostasis by reducing the amplitude of entry through CNG channels of Ca 2+ propagating the antiproliferative signal from the membrane to the nucleus.Tachyphylaxis induced by heat-stable enterotoxins was not altered by inhibition of PDE5 or application of 8-br-cGMP, suggesting that cGMP hydrolysis is not central to this process.Rather, 8-br-cGMP, a potent activator of PKG (44), induced tachyphylaxis to acute cGMP signaling.In addition, removal of extracellular Ca 2+ from the medium did not affect the development of tachyphylaxis to heat-stable enterotoxins, indicating that Ca 2+ entering through CNG channels and the associated downstream intracellular effectors did not mediate acute desensitization.These observations are particularly relevant, because the influx of extracellular Ca 2+ through store-operated Ca 2+ channels, but not through CNG channels, opposes heatstable enterotoxin-dependent cytostasis by allosterically inhibiting GCC-induced accumulation of [cGMP] i in human colon cancer cells (45).
GCC activation by heat-stable enterotoxins engages a previously unrecognized intracellular signaling network regulating cell cycle progression of colon carcinoma cells (Fig. 6).Cyclic GMPdependent activation of PKG and PDE5 represents one gating mechanism for this network, shaping the duration and amplitude of heat-stable enterotoxin-induced cytostasis.In this model, elevated [cGMP] i generated by GCC activation inhibits the tumor cell cycle by stimulating Ca 2+ entry through CNG channels (13).Acutely, cGMP signaling induces PKG-dependent reduction of Ca 2+ entry through CNG channels, uncoupling cytostasis from [cGMP] i .Subsequently, prolonged GCC stimulation induces cGMP-dependent activation of PDE5 (25,27), which lowers [cGMP] i , further desensitizing colon cancer cells to heatstable enterotoxin actions.PKG likely plays a role in tachyphylaxis and bradyphylaxis, by regulating Ca 2+ entry through CNG channels and sensitizing PDE5 to allosteric cGMP stimulation (26,39).Notably, elimination of negative feedback mechanisms in cGMP signaling with specific pharmacologic inhibitors of PDE5 and PKG completely rescues heat-stable enterotoxininduced cytostasis.
Mechanisms underlying desensitization to GCC activation described herein may have important implications for intestinal homeostasis.Indeed, the endogenous GCC ligands, guanylin and uroguanylin, are novel regulators of intestinal epithelial cell dynamics.Uroguanylin mimics the effects of heat-stable enterotoxins on intestinal epithelial cell proliferation (15) and inhibits  1A-C .ZAP , 10 Amol/L zaprinast; RP8pCPT , 100 Amol/L RP8pCPT-cGMPs.For DNA synthesis (A), cells were preincubated (20 hours) with PBS, and 1 Amol/L heat-stable enterotoxins (ST, left ) or 5 mmol/L 8-br-cGMP (right ).Zaprinast was added 15 minutes before the final application of heat-stable enterotoxins, whereas RP8pCPT-cGMPs was added during preincubations.For cell proliferation (B) and flow cytometry (C ), treatments were conducted during the period of stimulation with serum (24 hours).*, P < 0.05; **, P < 0.01, compared with effects of heat-stable enterotoxins alone (A, left and B) or 8-br-cGMP alone (A, right ) in desensitized cells.
polyp formation in an animal model of colon cancer ( 14), whereas guanylin-null mice exhibit increased rates of colonocyte proliferation (46).Thus, stimulation of GCC signaling by endogenous ligands and the associated negative feedback mechanisms regulating cGMP signaling may be central to the transition from proliferation to differentiation along the cryptvillus axis.Furthermore, homologous desensitization may be protective in the context of infections by enterotoxigenic bacteria, which produce secretory diarrhea that is characteristically self-limited by incompletely characterized mechanisms (8,21).Conversely, the reversibility of desensitization maintains the sensitivity of cGMP-dependent cytostatic mechanisms to intermittent longitudinal exposure to heat-stable enterotoxins.Indeed, cGMP-dependent cytostasis, intermittent exposure to heat-stable enterotoxins, and the reversibility of desensitization may, in part, underlie the resistance to colon cancer of populations in geographic areas with endemic enterotoxigenic Escherichia coli (13,14).
In conclusion, cytostasis induced by GCC in human colon cancer cells is limited by cGMP-dependent homologous desensitization mediated by PKG-and PDE5-induced tachyphylaxis and bradyphylaxis, respectively.These observations underscore the plasticity of cancer cells to develop adaptive strategies that evade antiproliferative signals by reducing intracellular concentrations of key intermediates, a common phenomenon characterizing multidrug resistance to anticancer agents (5,35).Importantly, disruption of cGMP-dependent tachyphylaxis and bradyphylaxis completely restores heat-stable enterotoxin-induced cytostasis of human colon cancer cells.Of significance, pharmacologic inhibition of PDE5, a principle determinant of cGMP-dependent bradyphylaxis induced by heat-stable enterotoxins, is extensively employed in managing patients with erectile dysfunction (47) and pulmonary hypertension (48).The availability of those agents for in vivo use should facilitate future studies examining the use of cGMP-targeted combination therapy incorporating inhibitors of cGMP-dependent desensitization and cytostatic bacterial enterotoxins for the prevention and treatment of colorectal cancer.

Figure 6 .
Figure 6.Proposed mechanisms underlying negative feedback regulation of heat-stable enterotoxin-induced cytostasis in human colon cancer cells.