Enhancing or Eliminating Signals for Cell Survival to Treat Disease

Abstract

Approximately half of all medical illnesses can be attributed to insufficient or excessive apoptosis. Apoptosis resistance is a cardinal feature of cancer, mediated in many instances, by signal transducer and activator of transcription (STAT) 3. We identified G-quartet oligodeoxynucleotides (GQ-ODNs) as potent and selective inhibitors of Stat3 DNA binding activity in vitro. We report here that GQ-ODNs are capable of inhibiting the growth of nude mouse xenografts of breast and prostate tumors. We developed a rat model of severe hemorrhagic shock (HS) to assess the benefits of promoting Stat3 activity in diseases marked by excessive apoptosis. Administration of the Stat3-activating cytokine IL-6 at the initiation of resuscitation from HS activated intra-cardiac Stat3, reversed cardiac apoptosis, left ventricular dysfunction and hypovolemic circulatory collapse (HCC) and resulted in a 5-fold reduction in mortality; pre-treatment of rats with GQ-ODN prevented the reversal of cardiac apoptosis and HCC by IL-6. Thus, targeting of Stat3 may be a useful for treatment of multiple cancers; agents that activate Stat3 may be beneficial in acute insults that cause apoptosis in organs critical for survival.

Introduction

The adult human body contains approximately 10¹⁴ cells, a number maintained in health by a balance between mitosis and programmed cell death or apoptosis (1). Approximately half of all medical illnesses can be attributed to insufficient or excessive apoptosis (1). Diseases with too little apoptosis include cancer and autoimmune disorders; diseases with too much apoptosis include stroke, myocardial infarction, hemorrhagic shock, toxin exposure, AIDS and neurodegenerative diseases. For many of these diseases, adequate therapy is either lacking or suboptimal.

Recently, our strategy for developing new treatments for diseases marked by decreased or increased apoptosis has been to target or promote, respectively, the activity of signal transducer and activator of transcription (STAT) 3 (Figure 1A). Stat3 is a latent transcription factor present in the cell cytoplasm in a head-to-head inactive dimer configuration (2). Stat3 is activated by stimulation of cells with cytokines such as interleukin (IL)-6, which results in phosphorylation of Stat3 on tyrosine (Y) 705 leading to reconfiguration of Stat3 into tail-to-tail dimers, a conversion mediated by reciprocal interactions between the pY705 motif within one Stat3 proteins and the Stat3 Src homology (SH) 2 domain of its partner (Figure 1C). This reconfiguration promotes Stat3 translocation to the nucleus and facilitates binding to specific sites within the genome including promoters for anti-apoptosis genes resulting in activation of gene transcription.

Fig. 1.

Models of Stat3, GQ-ODN and IL-6 signaling. Panel A is a schematic of Stat3 showing its structural domains and their amino-acid-residue boundaries including the N-terminal tetramerization domain (responsible for constitutive head-to-head inactive dimer formation within the cytoplasm and for tetramerization within the nucleus at sites of duplicate Stat3-binding promoter sites each of which binds an activated Stat3 dimer), the coiled-coil (responsible for Stat3-protein interactions) the DNA binding domain, the linker domain, the Src-homology (SH) 2 domain (responsible for tail-to-tail active dimer formation through reciprocal SH2-pY705 motif interactions) and the trans-activation domain (responsible for optimal activation of gene transcription). The key phosphorylation site Y705 is also indicated. Panel B depicts the sequence of GQ-ODN T40214 and its K⁺-stabilized structure based on NMR analysis. The K⁺ ions are depicted in the grey circles. Panel C summarizes the main pathways and proteins involved in IL-6 signaling including IL-6, IL-6Rα and β chains and Janus kinases (JAKs) common to all pathways. The Src homology protein tyrosine phosphatase (SHP)-2 is recruited to the pY759 motif within the IL-6Rβ chain. It recruits Grb2 and Ras proteins that activate p38 mitogen activated protein kinase (MAPK) and phosphoinositol-3 kinase (PI3K). These pathways diverge with PI3K binding to and activating Akt kinase, which phosphorylates the transcription factors Forkhead (FKD), nuclear factor (NF)-κB and cAMP response element binding protein (CREB) which translocate to the nucleus where they can bind to the promoters of anti-apoptotic genes. Also shown are the YXXQ motifs within the IL-6Rβ chain that when phosphorylated can recruit inactive head-to-head Stat3 dimers, resulting in phosphorylation of Stat3 Y705 and reconfiguration into active tail-to-tail dimers, which translocate to the nucleus where they can bind to the promoters of anti-apoptotic genes. GQ-ODN is shown as a grey triple-decker cube structure that interacts with active Stat3 dimers inhibiting their ability to bind DNA.

Apoptosis resistance is a cardinal feature of cancer, one that has been attributed, in many instances, to constitutive activation of Stat3 (3). The mechanism for constitutive activation in some cases, including head and neck and prostate cancer, involves autocrine production of cytokines or growth factors such as IL-6 and TGFα; the mechanism for constitutive activation of Stat3 in other instances is not currently understood.

Increased apoptosis within multiple organ systems has been observed in ischemia/reperfusion injuries including hemorrhagic shock (4–11) but not within the heart. IL-6 has been demonstrated to reduce apoptosis following toxic insults (4,12–17). However, it is not known whether or not IL-6 administration prevents apoptosis following hemorrhagic shock and, if so, whether or not apoptosis prevention is mediated by Stat3.

We recently developed G-rich oligodeoxynucleotides, which form intramolecular G-quartet structures as a new class of Stat3 inhibitor for use as cancer therapy and in chemical genetics studies (Figure 1B and C). G-quartet oligodeoxynucleotides (GQ-ODN) were demonstrated to be potent and selective inhibitors of Stat3 DNA binding activity in vitro (18–20). Computer-simulated docking studies indicated that GQ-ODN mainly interacted with the SH2 domain of Stat3 and were capable of inserting between the SH2 domains of Stat3 dimers. In vitro DNA binding studies indicated that GQ-ODN destabilized Stat3 dimers thereby reducing their DNA binding activity.

We demonstrate here that administration of GQ-ODN intravenously was able to inhibit the growth in nude mice of xenografts of breast and prostate cancers in which Stat3 is constitutively activated. We also demonstrated that hemorrhagic shock-induced cardiac apoptosis, left ventricular dysfunction and mortality can be reversed by IL-6 administration; use of GQ-ODN in a chemical genetics approach revealed that the beneficial effects of IL-6 are mediated, in part, by activation of intracardiac Stat3. Thus, Stat3 warrants further consideration as a drug target for inhibition in diseases such as cancer marked by apoptosis resistance and for activation in diseases characterized by excessive apoptosis.

Materials and Methods

Reagents and chemicals.

GQ-ODN T40214 was synthesized by Midland Certified Reagent Company (Midland, Texas) and used without further chemical modifications. The human breast cancer cell line, MDA-MB-468, and the human prostate cell line, PC-3, were obtained from the ATCC. Polyethylenimine (PEI, ∼25 kD branched polymer, Aldrich Chemical, WI) was generously provided by Dr. Charles Densmore (Baylor College of Medicine). Human recombinant interleukin 6 (IL-6) produced in E. coli was the generous gift of Novartis.

Animal models.

These studies were approved by the Baylor College of Medicine Institutional Review Board for animal experimentation and conform to National Institutes of Health guidelines for the care and use of laboratory animals.

Nude mouse tumor xenograft model.

Athymic nude mice (Balb/nu/nu, four weeks old weighing ∼20g) were obtained from Charles River Labs. Mice were injected subcutaneously into the right or left flank with one million cancer cells (MDA-MD-468 or PC-3) in 200 μl PBS. After tumors were established at 7 to 14 days post-injection, nude mice were randomly assigned into two groups. Mice in Group 1 served as placebo and received only PEI (2.5 mg/kg) while mice in Group 2 received T40214 (5.0 mg/kg) plus PEI (2.5 mg/kg). Treatments, administered by tail vein injection, and sizing of tumors occurred every two days. The unpaired two-sample t-test, [t = (X₁ − X₂)/[S_{p²(1/n1 + 1/n2))^1/2}], was used to determine differences in tumor sizes between the placebo and the drug-treated groups.

Rat protocols for trauma plus hemorrhagic shock.

Adult male Sprague-Dawley rats were obtained from Harlan (Indianapolis, IN), and were randomly subjected to the sham protocol or one of two hemorrhagic shock (HS) protocols, reversible (R), HS or irreversible (I) HS (Figure 2). Animals were anesthetized with 2% isoflurane in room air administered through a nose cone. Both superficial femoral arteries were surgically prepared and cannulated with PE-50 tubing. The right superficial femoral catheter site was used for continuous blood pressure monitoring and the left superficial femoral catheter site was used for blood withdrawal or fluid administration. After an initial bleed of 2.25 mL/100 g body weight over 10 min, blood was withdrawn into a heparinized syringe episodically until compensation failed and was then returned as needed to maintain a mean arterial blood pressure (MAP) of 35 mmHg. At the point in time at which 35% (RHS) or 50% (IHS) of shed blood had been returned, the animals were resuscitated by administration of the remaining shed blood plus two times the total shed blood volume with lactated Ringer’s solution (LR; Baxter Laboratories, Deerfield, IL). For experiments described here, the rats were randomly subjected to RHS or IHS in two experimental groups. The first group, placebo group (P), received 0.1 ml phosphate buffered saline (PBS; Gibco, Invitrogen Corporation, Grand Island, N.Y.) at the initiation of the resuscitation. The second group, IL-6 group, received 10 μg/kg of recombinant human IL-6 in 0.1 ml PBS at the initiation of the resuscitation. Total shed blood amounts and shock times were monitored in the HS groups and were statistically equivalent. Sham animals underwent cannulation and anesthesia for an identical period of time as shock animals but were not bled or resuscitated. IHS animals were euthanized one hour after the initiation of resuscitation by exsanguination while still anesthetized; sham animals were euthanized at the corresponding time point. Hearts were immediately harvested and immersed in relaxation solution (21), cut horizontally into 3 equal-sized pieces; each portion was snap frozen in liquid nitrogen. RHS animals were monitored under anesthesia for 4 hr after the initiation of resuscitation at which point catheters were removed, incision sites sutured and animals allowed to awaken. Rats were observed over the next 48 hr; morbid animals were humanely euthanized and counted as non-survivors. Rats surviving after 48 hr were counted as survivors based on previous experience that mortality from HS occurs in the first 48 hr.

Fig. 2.

Idealized mean artery pressure (MAP) tracings for models of reversible and irreversible hemorrhagic shock (RHS and IHS, respectively). MAP is targeted for 35 Torr in both models. Rats are maintained at this MAP initially by withdrawing then by reinfusing small amounts of blood for approximately 3 hr in RHS and for approximately 4.5 hr in IHS before resuscitation as indicated. The gray region depicts the mean ± SD MAP of 107 normal rats. The percent resuscitation success, defined as return of MAP to within the gray area, and the percent mortality is shown for the RHS and IHS protocols.

To achieve pharmacological inhibition of Stat3 activity within the heart, rats were randomized to receive the G-quartet oligodeoxynucleotide (GQ-ODN) T40214 or non-specific (NS)-ODN (2.5 mg ODN/kg) complexed in polyethyleneimine by tail vein injection, as described (19), 24 hours prior to subjecting them to IHS protocol. Fluorescently labeled T40214 when delivered in this fashion was previously shown to accumulate within the cells of a variety of normal tissues including heart, liver and kidneys (Jing N, unpublished data 2003) as well as tumor tissue (19); half-life of T40214 within tumors was ≥48 hr (20).

Aorta Doppler ultrasound studies.

A 20 MHz Doppler ultrasound probe custom-built, as described (22), was used for these studies. Rats were subjected to Doppler evaluation, 30 min after the end of resuscitation, as described (23). Briefly, rats were laid in the supine position and held in place with tape. The fur on the chest was moistened with water to provide acoustic coupling and the probe placed at the second intercostal space immediately to the right of the sternum and angled toward the heart with the range gate set between 2 and 4 mm to obtain signals from the ascending aorta just above the valve. The quadrature audio signals from the Doppler probe were connected to a fast Fourier transform (FFT) spectrum analyzer and recorded on a high-fidelity videocassette recorder (VCR). Once acceptable signals were obtained, 10–20 s of audio signals were recorded on the VCR. The recorded signals were played back at actual speed into a zero-crossing-interval histogram (ZCIH) processor and 1 s of data was stored to a computer file at 37 kHz sampling rate for display and analysis. From the display, we measured period in milliseconds, ejection time in milliseconds and acceleration in cm per second squared.

ELISA detection of cytoplasmic nucleosomes.

Determination of cytoplasmic nucleosomes or histone-associated DNA fragments was performed using the Cell Death Detection ELISA Plus Kit (Roche), following the instructions of the manufacturer but modified for detection of nucleosomes within tissues. Briefly, portions of frozen hearts were cut into 4 micron sections using a cryostat microtome and sonicated briefly at 4° C in lysis buffer. The protein concentration of lysates were determined by Bradford assay and 200 μg protein added in duplicate to streptavidin-coated wells of 96-well microtiter plates, to which was added a mixture of anti-histone-biotin and anti-DNA-peroxidase antibody. Following a 2-h incubation and washing, ABTS (2,2′-azino-di[3-ethylbenzthiazolin-sulfonate]) was added. After 20 min, absorbance at 405 nm and 490 nm were determined. The absorbance at 405 was corrected by subtracting the absorption at 490 and the units of nucleosomes calculated using a dilution curve of a positive control provided by the manufacturer.

Terminal deoxynucleotidyl transferase-mediated dUTP nick end-labeling (TUNEL) staining.

Terminal deoxynucleotidyl transferase-mediated dUTP nick end labeling (TUNEL) was performed using the ApopTag® Plus Peroxidase In Situ Apoptosis Detection Kit (CHEMICON International, Inc., Temecula, CA) following the manufacturer’s instruction. One portion of the frozen heart was fixed and sectioned for staining. TUNEL-stained slides were counterstained with hematoxylin. TUNEL-positive nuclei were enumerated within twenty 1,000× magnified fields per slide by an experienced histologist blinded to the treatment the rat received.

Electrophoretic mobility shift assay (EMSA).

EMSA was performed using heart tissue extracts from experimental groups as described (24). Binding reactions were performed using 20 μg of extracted protein and radiolabeled high-affinity serum-inducible element (hSIE) duplex oligonucleotide that preferentially binds Stat3 and Stat1 (25). Binding reactions were separated on a 4–6% polyacrylamide gel, as described (26). The level of transcription factor activation was quantitated using PhosphorImager analysis combining the intensities of the Stat3 homodimer and Stat3/Stat1 heterodimer gel shift bands.

Statistical analysis.

Unless otherwise indicated, one-way ANOVA was used to evaluate for significant differences within more than two groups, followed by Student-Newman-Keuls to identify where the significant differences occurred. Student’s t-test was used to evaluate for significant difference between two groups and Fisher’s exact test to compare ratios.

Results

Intravenous administration of G-quartet ODN results in inhibition of growth of xenografts of tumors in nude mice by induction of tumor cell apoptosis.

To determine if targeting Stat3 within cancer cells can reduce tumor growth, nude mice were injected subcutaneously in the flank with the breast tumor cell line, MDA-MB-468, or the prostate tumor cell line, PC-3, both of which demonstrated constitutive Stat3 activity. Once tumors were measurable, mice were treated intravenously by tail vein injection with the GQ-ODN T40214 (5.0 mg/kg), shown previously to block Stat3 DNA binding (18) plus PEI (2.5 mg/kg) or PEI (2.5 mg/kg) alone as placebo every other day. The mean volume of breast tumor xenografts of placebo-treated mice increased 7-fold over 18 days while the mean sizes of tumors in T40214-treated mice remained unchanged over this time period (Table 1). Expressed another way, the growth rate of breast tumors of placebo-treated mice was 11% per day while the growth rate of T40214-treated breast tumors was −0.4% per day (p=0.001). The mean size of the prostate tumor xenografts of placebo-treated mice increased 9 fold over 10 days while the mean sizes of prostate tumors from T40214-treated mice increased only 2.2-fold, respectively (Table 2); the growth rate of prostate tumors from placebo-treated mice was 21% per day while the growth of tumors from T40214-treated mice was 9.1%/day (p=0.001).

TABLE 1.

Volume of MDA-MB-468 breast tumor xenografts in nude mice during treatment with PEI alone (Placebo) or GQ-ODN plus PEI (T40214)

Treatment (number)	Day 0 Size (mm3)	Day 2 Size (mm3)	Day 4 Size (mm3)	Day 6 Size (mm3)	Day 8 Size (mm3)	Day 10 Size (mm3)	Day 12 Size (mm3)	Day 14 Size (mm3)	Day 16 Size (mm3)	Day 18 Size (mm3)	ΔS = (S18-S0)/S0 ×100%
Placebo (I)	37.5	60	65	71.5	82.5	120	140	150	202.5	229.5	+512%
Placebo (II)	27	37.5	50	61	71.5	105	114	136	156	182	+574%
Placebo (III)	30	37.5	66	72	78	78	97.5	183.8	196	196	+533%
Placebo (IV)	16	27	37.5	45.5	60.5	60	65	89.4	105	105	+556%
Placebo (V)	17.5	17.5	20	30	40	45	77	126	126	147	+818%
Placebo (VI)	10	15	17.5	20	20.3	22.5	37.1	54	54	65	+550%
Sm ± RMSD	23 ± 3.7	32.4 ± 6.1	42.7 ± 7.9	50 ± 8.1	58.8 ± 8.7	71.8 ± 10.8	88.4 ± 13.6	123.2 ± 17	140 ± 21.1	154 ± 22.7	+570%
Variation %	100 ± 16	140 ± 26	186 ± 34	217 ± 35	256 ± 37	312 ± 46	384 ± 59	536 ± 60	609 ± 65	670 ± 65
T40214 (I)	66	60	41	41	50	48	54	54	54	65	−1%
T40214 (II)	40	32	30.4	37.5	37.1	40.5	50	50	75	75	+87%
T40214 (III)	60	60	55	45	31	34	30	30	30	30	−50%
T40214 (IV)	60	50	37.5	24	24	21	24	24	21	21	−65%
T40214 (V)	70	80	80	72	65	54	67.5	84.4	84.4	90	+28%
T40214 (VI)	24	30	17.5	17.5	15	15.8	23.6	22.5	22.5	28.8	+20%
T40214 (VII)	10	12	6	1.5	1.5	0	0	0	0	0	−100%
T40214 (VIII)	24	36	40.5	36	30	33	40.5	40.5	33	33	+37%
Sm ± RMSD	44.3 ± 7.5	45 ± 7.1	38.5 ± 7.4	34.3 ± 6.4	31.7 ± 6.5	30.8 ± 5.8	36.2 ± 7.0	38.2 ± 8.3	40 ± 9.5	42.9 ± 10.8	−2%
Variation %	100 ± 16	101 ± 16	87 ± 17	77 ± 14	72 ± 15	70 ± 13	82 ± 16	86 ± 19	90 ± 21	97 ± 24

S_m (= ΣS_i/n) is a mean value of tumor sizes.

RMSD (= [Σ(S_i-S_m)²]^1/2/n) is root mean square deviation.

Variation (%)=S_i/S_o × 100 ± RMSD_i/S_o × 100 (%); Variation of S_o=100%.

TABLE 2.

Volume of PC-3 prostate tumor xenografts in nude mice during treatment with PEI alone (Placebo) or GQ-ODN plus PEI (T40214)

Treatment of nude mice	Day 0 Size (mm3)	Day 2 Size (mm3)	Day 4 Size (mm3)	Day 6 Size (mm3)	Day 8 Size (mm3)	Day 10 Size (mm3)	Change = (S10-S0)/ S0 × 100%
Placebo (I)	40	70	80	146.3	189	300	+650%
Placebo (II)	8	11	45	85.5	100	115.5	+1343%
Placebo (III)	13.5	23.6	22.5	39	97.5	97.5	+622%
Placebo (IV)	12	15	49.5	91	157.5	168	+1300%
Placebo (V)	28	57.8	63	100	170	229.5	+720%
Placebo (VI)	21	30	74.3	99	172.5	318.5	+1417%
Sm ± RMSD	20.4 ± 4.5	34.6 ± 8.9	55.7 ± 7.8	93.5 ± 12.8	147.8 ± 14.6	204.8 ± 34.7	+904%
Variation %	100 ± 22	170 ± 23	273 ± 14	458 ± 14	721 ± 10	1004 ± 10
Drug treated (I)	156	121	137.5	207	253.5	324	+107%
Drug treated (II)	240	252	300	342	374	385	+60%
Drug treated (III)	132	198	204.8	231	195	325	+146%
Drug treated (IV)	82.5	70	54	144.4	175.5	232.8	+182%
Drug treated (V)	28	21	45	70	100	142.5	+213%
Drug treated (VI)	20.3	20.3	33	33	60	60	+196%
Sm ± RMSD	109.8 ± 31.1	113.7 ± 35.5	129.1 ± 39.7	171.2 ± 42.2	193 ± 41.8	244.9 ± 46.3	+123%
Variation %	100 ± 28	104 ± 30	118 ± 30	156 ± 24	176 ± 21	223 ± 19

S_m (= ΣS_i/n) is a mean value of tumor sizes.

RMSD (= [Σ(S_i-S_m)²]^1/2/n) is root mean square deviation.

Variation (%)=S_i/S_o × 100 ± RMSD_i/S_o × 100 (%); Variation of S_o=100%.

To gain insight into the mechanism of inhibition of tumor growth by GQ-ODN T40214, we harvested the prostate tumor xenografts from placebo-treated and drug-treated mice after five treatments to perform TUNEL analysis and to assess for levels of Stat3-pY705, Bcl-x_L, Bcl-2 and activated Caspase 3 protein. TUNEL staining was performed on 4 tumor samples from placebo-treated mice and 3 tumor samples from drug-treated mice. The percentage of apoptotic cells was increased nearly 8-fold in the tumors of drug-treated mice (83.6 ± 1.0%) compared to the tumors of placebo-treated mice (11.2 ± 10.1%. p < 0.001). Furthermore, levels of Stat3-pY705, Bcl-x_L and Bcl-2 were decreased by 9, 4.3 and 10-fold, respectively, in the tumors from drug-treated animals compared to tumors from placebo-treated mice. These changes were accompanied by a 3-fold increase in Caspase 3 cleavage products in the tumors from drug-treated animals compared to tumors from placebo-treated mice (data not shown). In separate studies, GQ-ODN T40214 demonstrated similar activity and mechanism of action against tumor xenografts of squamous cell carcinoma of the head and neck (20).

Development of a rat model of hypovolemic circulatory collapse (HCC); HCC is accompanied by left ventricular contractile dysfunction and cardiomyocyte apoptosis.

Trauma is the leading cause of death in the US in persons 1 to 44 years of age. Half of trauma deaths are attributable to hypovolemic circulatory collapse (HCC). To gain insight into the mechanisms of HCC at the organ, cellular and molecular level, in particular, the contribution of cardiac apoptosis to this pathologic process, we developed a rat protocol of trauma plus irreversible (I) HS to model HCC. The hemorrhagic shock procedure involves trauma (bilateral groin dissections to isolate the superficial femoral arteries and cannulation of both arteries) and IHS (initial bleed of 2.25 ml/100g body weight over 10 min from the left artery to achieve a target MAP of 35 mm Hg). If rats are maintained at a MAP of 35 mm Hg until 50% of the shed blood has been returned, only 33% of rats (Table 3) can be resuscitated successfully-defined as achieving a MAP at the end of resuscitation of ≥72 mm Hg, the mean (94 mm Hg) minus two SD (11 mm Hg) of starting BP for 107 normal male Sprague-Dawley rats. Thus, two-thirds of rats subjected to the IHS protocol experienced HCC.

TABLE 3.

Effect of IL-6 on percent resuscitation success, nucleosome ELISA and number of TUNEL-positive cardiomyocytes rats subjected to sham and IHS protocol

Endpoint	Sham (n)	IHS/Placebo (n)	IHS/IL-6 (n)	IHS/IL-6/GQ (n)	IHS/IL-6/NS (n)
Resuscitation Success (%)	ND*	33 (6)1	100 (7)1,2	12.5 (16)2,3	94 (14)3
Nucleosome levels (U/mg total protein)**	2.4 ± 0.2 (5)1	657 ± 73 (4)1,2	21 ± 17 (5)2,3	223 ± 62 (6)3,4	30 ± 15 (5)4
Number TUNEL-positive nuclei/1,000× field**	1.3 ± 0.2 (3)1	16.1 ± 2.0 (3)1,2	8.5 ± 0.2 (3)2,3	16.3 ± 1.7 (3)3,4	11.8 ± 1.4 (3)4

* ND=not done.

^1,2,3,4Statistically different, p < 0.05.

^**Mean ± SEM.

It has been demonstrated in large animal models of HS that cardiac dysfunction contributes to HCC (27). To determine if this was the case in our model, we performed Doppler ultrasound studies of the ascending aorta; peak aortic acceleration has been demonstrated previously to be a good noninvasive index of cardiac contractile function in animals (23,28). Peak aortic acceleration in the IHS/P rats (6,974 cm/s²) was reduced by 50% compared to sham rats (14,061 cm/s²; p < 0.05) indicating that HCC in our model was accompanied by left ventricular contractile dysfunction.

Cardiomyocyte apoptosis has been observed in other settings of ventricular contractile dysfunction such as following myocardial infarction (29–31). To assess if HCC is accompanied by cardiomyocyte apoptosis, we evaluated the hearts of rats subjected to IHS for apoptosis using two assays—nucleosome ELISA and TUNEL staining. Nucleosomes were undetectable in sham rat hearts, but were readily detected within lysates of ventricles from IHS/P rats (Table 3). Results of TUNEL staining confirmed these findings (Table 3) and indicated that cardiomyocytes were the major contributor to nucleosomes. The number of TUNEL-positive nuclei in the IHS/P hearts (16.2 ± 2.0 per 1,000× field) was increased 12 fold from Sham hearts (1.3 ± 0.2 per 1,000× field; p < 0.5); histological assessment of the TUNEL-positive nuclei indicated that most were characteristic of the nuclei of cardiomyocytes.

IL-6 reverses HCC, ventricular contractile dysfunction and cardiomyocyte apoptosis following trauma plus IHS.

IL-6 has previously been demonstrated to protect organs from injury following HS (32,33) and other animal models of organ injury (12,13,15,16). We investigated whether administration of IL-6 at the start of resuscitation in rats subjected to IHS could prevent HCC (Table 3). Compared to only 33% of IHS/P rats, 100% of rats that received IL-6 (IHS/IL-6 rats) were successfully resuscitated (p < 0.05, Fisher’s exact test).

To determine if prevention of HCC by IL-6 was a result of improved left ventricular contractile function, we determined peak aortic acceleration of IHS/IL-6 rats and compared the results with IHS/P rats. Resuscitation with IL-6 completely reversed the IHS-induced ventricular contractile dysfunction (p < 0.05, one-way ANOVA). To determine if cardiomyocyte apoptosis following trauma plus IHS is reversed by IL-6, we performed nucleosome ELISA assays (Table 3). Nucleosome levels in hearts from IHS/IL-6 rats were reduced 94% compared to IHS/P rats (p < 0.05). To verify these ELISA results, we performed TUNEL assays of sections of rat hearts (Table 3). The number of TUNEL-positive nuclei in the IHS/IL-6 rats (8.5 ± 0.2 per 1,000× field) was reduced by 47% compared to IHS/P (16.1 ± 2.0; p < 0.05).

IL-6 decreases mortality after trauma and HS.

Due to the severity of injury in the trauma plus IHS protocol (mean 4.5 hr of hypotension), all rats die within 2-to-4 hours regardless of treatment with IL-6 indicating that the IHS protocol is too severe to assess whether IL-6 treatment confers a survival benefit. To investigate the impact of IL-6 on mortality after trauma plus hemorrhage, we subjected the rats to a less severe HS protocol (RHS; mean 3 hr of hypotension) and randomized them to receive either placebo or IL-6 (10 μg/kg) at the start of resuscitation. Rats that received IL-6 at the start of resuscitation demonstrated a nearly 5-fold reduction in mortality (15.4%) compared to placebo-treated controls (72%; p < 0.0001, Fisher’s exact test).

The ability of IL-6 to prevent HCC and cardiomyocyte apoptosis is Stat3 dependent.

IL-6 activates Stat3, which has previously been demonstrated to activate the transcription of several anti-apoptotic genes and to contribute to apoptosis resistance in cancer cells (reviewed in (34)). To begin to assess if the anti-apoptotic effect of IL-6 is mediated by Stat3 activation, we first determined if Stat3 is activated in the hearts of rats resuscitated with IL-6. Extracts of cryotome sections of the heart harvested 1 hour after IL-6 treatment were examined by EMSA. Phosphoimaging analysis of the Stat3 bands indicated that Stat3 activity is increased over two fold in the hearts of IHS/IL-6 rats compared to IHS/P rats (p=0.028, Student’s t-test).

To determine if activation of Stat3 downstream of IL-6 was important for the improved resuscitation observed in IHS/IL-6 rats, we pre-treated rats 24 hr before IHS and IL-6 resuscitation with a G-quartet-oligonucleotide (T40214) that specifically binds to and inhibits the activity of Stat3 (19) or with a non-specific ODN. Stat3 activity was reduced 3-fold within the hearts of IHS/IL-6/GQ-ODN compared to IHS/IL-6/NS-ODN rats (data not shown). Resuscitation was successful in only 13% of IHS/IL-6/GQ-ODN rats while resuscitation was successful in 94% of IHS/IL-6/NS-ODN rats (p < 0.05, Fisher’s exact test; Table 3). To determine if IL-6 activation of Stat3 was also important for the anti-apoptotic effects of IL-6 in IHS, we examined the hearts of IHS/IL-6 rats pretreated with GQ-ODN vs. NS-ODN for nucleosome levels and TUNEL-positive nuclei. Nucleosome levels in the hearts of IHS/IL-6/GQ-ODN rats were increased 10-fold compared to hearts from IHS/IL-6 rats (p < 0.05) to within 34% of IHS/P rats; the number of TUNEL-positive nuclei was similar to that of IHS/P rats (Table 3). In contrast, nucleosome levels in the hearts of IHS/IL-6 rats pre-treated with NS-ODN were identical to the levels in the IHS/IL-6 group (Table 3); the results of TUNEL staining in IHS/IL-6 rats pre-treated with NS-ODN (Table 3) were consistent with the nucleosome results. Together these results indicated that Stat3 mediates a substantial portion of the effect of IL-6 on inhibiting cardiomyocyte apoptosis.

Discussion

Our strategy for developing new treatments for diseases marked by decreased or increased apoptosis recently has been to target or promote, respectively, the activity of signal transducer and activator of transcription (STAT) 3. We developed G-rich oligodeoxynucleotides, which form intramolecular G-quartet structures as a new class of Stat3 inhibitor. G-quartet oligodeoxynucleotides (GQ-ODN) were demonstrated to be potent and selective inhibitors of Stat3 DNA binding activity in vitro (18). Intravenous injection of GQ-ODN complexed with polyethyleneimine (PEI) into nude mice with breast and prostate tumor xenografts markedly reduced tumor growth compared to mice treated with PEI alone; reduction in tumor growth was accompanied by increased numbers of apoptotic cells within tumors and reduction in Stat3 activation and levels of classical anti-apoptotic proteins. To assess the benefits of promoting Stat3 activity in disease states marked by excessive apoptosis, we developed a rat model of trauma and severe hemorrhagic shock (HS) that mimics hypovolemic circulatory collapse (HCC). HCC is the cause of half of all deaths from trauma, which is the leading cause of death for individuals in the US between the ages of 1 and 44. Our findings in this model demonstrated that cardiomyocyte apoptosis is a crucial contributor to the development of HCC and cardiovascular dysfunction in HS. Administration of IL-6 (10 μg/kg) at the initiation of resuscitation from HS activated intra-cardiac Stat3, reversed HCC, cardiovascular dysfunction and cardiac apoptosis, as well as resulted in a 5-fold reduction in mortality. Pre-treatment of rats with GQ-ODN plus PEI but not with a non-specific ODN, blocked the reversal of HCC and cardiac apoptosis. Thus, targeting of Stat3 with agents such as GQ-ODN may be a useful strategy for treatment of multiple cancers, while agents such as IL-6 that activate Stat3 merit consideration for intervention in acute insults that cause apoptosis in organs critical for survival.

Our previous studies showed that the G-quartet structure of G-rich ODN is essential for the inhibition of Stat3 DNA-binding activity in vitro (18). The G-quartet structure of T40214 closely resembles a perfect cylinder 15 Å in width and length. This conformation increases the thermal stability of the structure, reduces the capacity of ODN to form molecular aggregates and prevents single-strand endonucleases from accessing their cleavage sites (35).

Effective delivery of GQ-ODNs into cells is key for success of Stat3 inhibitors to be effective anti-cancer and chemical genetic agents. A novel intracellular delivery system has been developed for GQ-ODNs that is based on the requirement for potassium for G-rich ODN to form the G-quartet structures within cells under physiological conditions (36). Using the novel delivery system, GQ-ODNs were delivered efficiently into the cytoplasm and nucleus of cells. When G-rich ODNs form intramolecular G-quartet structures within the cytoplasm, they are able to diffuse freely through pores into the nucleus, bind to Stat3 and block the transcription of Stat3 regulated genes.

Intravenous administration of GQ-ODN plus PEI was well tolerated by mice at the dose used in this study. Detailed toxicity studies of T40214 are in progress; however, toxicity studies of GQ-ODN T30177, an analog of T40214 that inhibits HIV-1 integrase, has been performed previously (37). T30177 did not induce genetic mutations in three assays—the Ames/Salmonella mutagenesis assay, the CHO/HGPRT mammalian cell mutagenesis assay and the mouse micronucleus assay. Acute toxicity studies in mice revealed an LD₅₀ for T30177 of 1.5 g/kg body weight; chronic toxicity studies in mice following multiple doses did not cause delayed mortality or changes in serum chemistry, hematological parameters or organ histology until the dose of T30177 reached 600 mg/kg, 120 times the dose (5 mg/kg) used in our studies.

While left ventricular dysfunction in severe HS and its physiological details were reported nearly half a century ago (27), its cellular and molecular basis has remained incompletely defined. Cardiomyocyte apoptosis has been well described in other cardiac insults including rodent coronary artery occlusion models (29–31). Furthermore, inhibition of caspase activity within the myocardium (30) or transgenic over-expression of the anti-apoptotic protein Bcl-2 within cardiomyocytes (31) in these models resulted in decreased cardiac apoptosis and improvement in ventricular function. While apoptosis has been demonstrated in multiple organs after HS (4–11), it has not been previously demonstrated to occur within the heart following this insult. We observed a dramatic increase in cardiomyocyte apoptosis as measured by nucleosome ELISA and TUNEL in the hearts of rats subjected to our IHS protocol that results in HCC; IL-6 treatment reversed HCC, left ventricular dysfunction and cardiomyocyte apoptosis strongly supporting the hypotheses that cardiomyocyte apoptosis contributes to HCC and left ventricular dysfunction following IHS and that agents such as IL-6, which prevent cardiomyocyte apoptosis, can be beneficial in this setting.

Binding of IL-6 to its receptor induces oligomerization and phosphorylation of the IL-6 receptor β chain, glycoprotein (gp) 130 (reviewed in (38)). Phosphorylation of gp130 leads to recruitment and activation of intermediates involved in three signaling pathways that effect apoptosis—Ras mitogen-activated protein kinase (MAPK), phosphatidylinositol 3-kinase-dependent (PI-3K), and Stat3—the last two of which link to cell survival. Activation of PI-3K activates Akt, which is involved in multiple anti-apoptotic mechanisms, including phosphorylation of the Bcl-2 family members and activation of transcription factors—NF-κB, Forkhead and cyclic adenosine 3′,5′-monophosphate response element binding protein (CREB)—resulting in increased transcription of survival genes (39). IL-6 activation of Stat3 leads to up-regulation of anti-apoptosis genes such as Bcl-xL, Bcl-2, Mcl-1 and FLIP (40–43). In vitro and in vivo findings of others have provided evidence that Stat3 protects against cardiomyocyte apoptosis in some cardiac insults including ischemia-reperfusion and toxin-mediated injuries (44–46). However, the role of cardiac Stat3 in HS has not been investigated. We demonstrated a 2-fold increase in Stat3 activity within the hearts of rats that received IL-6 compared to placebo-treated rats. Furthermore, we demonstrated that pre-treatment of rats with a specific GQ-ODN inhibitor of Stat3 blocked the IL-6-mediated increase in intra-cardiac Stat3 along with the ability of IL-6 to prevent HCC and to inhibit cardiomyocyte apoptosis. These findings indicate that the cardioprotective effects of IL-6 are mediated, at least in part, by Stat3.

DISCUSSION

Wolf: Boston: What happens when you give IL-6 to an animal whose coronaries have been occluded and then reopened?

Tweardy: Houston: It’s anyone’s guess, although I suspect that the effects would be similar to what we observed in hemorrhagic shock, since many of the pathophysiological pathways are similar; although the level of insult is probably greater than in hemorrhagic shock. It’s something we are interested in investigating, as well as the effect of IL-6 administration following cardiac arrest.

Wolf: Have you done that yet?

Tweardy: We have not done that, but it’s a very good idea.

Bast: Houston: The effect of IL-6 suggests that there could be IL-6 receptors on cardiac myocytes. Alternatively, there could be an indirect effect. Do you have any evidence one way or another on that?

Tweardy: In normal heart, there’s no convincing evidence yet for expression of the IL-6 receptor alpha chain. There certainly is expression of gp130, which is the IL-6 receptor beta chain. The IL-6 receptor is a heterodimeric complex of these two proteins. The question you ask is a good one. We think that early in resuscitated hemorrhagic shock and perhaps in other ischemia/reperfusion injuries, there is generation of increased levels of soluble IL-6 receptor alpha from cells that express membrane-bound IL-6 receptor alpha. The presence of soluble IL-6 receptor in the circulation makes virtually every tissue in the body that expresses gp130, which is essentially all tissues, now responsive to IL-6.

Thorner: Charlottesville: That was a beautiful presentation. There is cross-talk between the STAT; molecules including Stat3 and there are multiple hormones and cytokines that activate them. So is this really selective IL-6 or is this just the beginning of a whole big story?

Tweardy: You are absolutely correct. There are a number of growth factors, cytokines, and other hormone-like molecules that signal through Stat3 and it’s certainly possible that IL-6, while being one ligand that can activate Stat3, may not be the best ligand. The question of what is the best way to activate Stat3 remains to be answered.

Billings: Baton Rouge: Dave, I was wondering, your being an infectious disease fox, if you had looked at both Gram-negative and Gram-positive shock and sepsis with your IL-6?

Tweardy: We have not. It is a more difficult question to address since, as you know, sepsis doesn’t have a clear-cut time of initiation that is amenable to intervention. The reason I was attracted to hemorrhagic shock is that much of what we see, similar to ischemia/reperfusion injuries in terms of organ and inflammation, requires the resuscitation phase. This allowed us not only to investigate what was going on but to think therapeutically as well. There’s a lot of evidence that you are probably aware of that apoptosis occurs following sepsis, and may predispose to secondary infectious complications in septic patients. We actually have looked at this issue in hemorrhagic shock in rodent models and are beginning to look at this issue in shock/trauma patient populations. In a mouse model of hemorrhagic shock, IL-6 prevented liver apoptosis and reduced the burden of organisms in the liver following intra-peritoneal inoculation of Staphylococcus aureus. This one study indicates that there is a benefit in reducing apoptosis following hemorrhagic shock with IL-6 in terms of reducing subsequent susceptibility to infection.

Boxer: Ann Arbor: Two questions David. The first, once apoptosis begins can IL-6 treatment reverse it, and 2) what impact does Stat3 activation have on stabilizing myodochondrial membranes?

Tweardy: Those are two good questions. Your first question can be restated as: “Is there a window during which the anti-apoptotic effect of IL-6 effect can be observed?” We actually have not investigated this yet. What we have done so far is give IL-6 at the initiation of resuscitation, so the very first thing that goes into the veins, once we begin the resuscitation is IL-6. We have not tried to see how long we can wait to give IL-6 to still see the benefit. So that’s still an open question. At the core of your second question is how does Stat3 work; specifically does it work by stabilizing mitochondrial membranes? We postulated that activated Stat3 was up-regulating anti-apoptotic members of the Bcl2 family leading to resistance to mitochondrial leakage. However, thus far, we have not found increases in their levels of mRNA expression. In fact, while the mRNA levels of several of the members of the Bcl2 family increased with hemorrhagic shock, their levels normalized in the shock animals treated with IL-6. As I very quickly alluded to during the talk, our current thinking is that Stat3 may be functioning as a negative transcriptional regulator in the hearts of our IL-6-treated rats.