| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Human Cancer Biology |
Authors' Affiliations: Departments of 1 Cell Stress Biology and Photodynamic Therapy Center, 2 Molecular and Cellular Biology, and 3 Medicine, Roswell Park Cancer Institute, Buffalo, New York
Requests for reprints: Heinz Baumann, Department of Molecular and Cellular Biology, Roswell Park Cancer Institute, Elm and Carlton Streets, Buffalo, NY 14263. Fax: 716-845-5908; E-mail: Heinz.baumann{at}roswellpark.org.
 |
Abstract |
|---|
 Top Abstract Materials and MethodsResultsDiscussionReferences |
|---|
Experimental Design: Cells in tissue culture, murine s.c. tumors, and endobronchial carcinomas in patients were treated with PDT, and the noncleavable cross-linking of the latent signal transducer and activator of transcription 3 (STAT3) was determined.
Results: Murine and human cancer cell lines reacted to PDT by an immediate covalent cross-linking of STAT3 to homodimeric and other complexes. The magnitude of this effect was strictly a function of the PDT reaction that is determined by the photosensitizer concentration and light dose. The cross-link reaction of STAT3 was proportional to the subsequent cytotoxic outcome of PDT. An equivalent photoreaction as detected in vitro occurred in tumors treated in situ with PDT. The light dose-dependent STAT3 cross-linking indicated the relative effectiveness of PDT as a function of the distance of the tissue to the treating laser light source. Absence of cross-links correlated with treatment failure.
Conclusions: The data suggest that the relative amount of cross-linked STAT3 predicts the probability for beneficial outcome, whereas absence of cross-links predicts treatment failure. Determination of STAT3 cross-links after PDT might be clinically useful for early assessment of PDT response.
The range of PDT-dependent modifications of cellular proteins includes oxidation, adduct formation, protein cross-linking, and degradation (5–9). Protein cross-linking, long known to occur as a consequence of PDT (2), has recently drawn new attention with the identification of highly specific substrates for the cross-linking reaction, including important cell signaling molecules, such as epidermal growth factor receptor (10, 11), Bcl-2 (3), and members of the nonactivated signal transducer and activator of transcription (STAT) family (11).
PDT-mediated STAT cross-linking has the following characteristics (11): (a) PDT induces cross-linking of preformed homodimeric complexes of nonactivated cytoplasmic STAT3 and, to a lesser extent, of STAT1 and STAT4; (b) STAT cross-linking is a function of the PDT photoreaction, occurring only during the light treatment period, and correlates with PDT dose; (c) cross-linked STAT proteins are stable (half-life of 
18 h) and detectable immediately after completion of treatment; (d) STAT cross-linking is observed with all cell types and photosensitizers tested thus far; and (e) STAT cross-links can be detected in PDT-exposed tissue in vivo. It has to be noted that (a) the PDT-mediated cross-link reaction does not involve STAT3 that has been activated by phosphorylation at Tyr705; (b) PDT cross-linked STAT3 dimer is similar in electrophoretic mobility to that generated by oxidative disulfide formation involving Cys259 (12) but differs in that the linking is noncleavable by reduction; and (c) cross-linked STAT3 is incapable of transducing cytokine and growth factor signals, but is not causative for PDT-induced cell death, as shown by PDT killing of human prostate PC3 cells that lack STAT3.
The above characteristics of STAT (particularly STAT3) cross-linking raise the possibility that it may represent a quantitative marker for the immediate, direct cellular PDT effect and may indicate the relative level of post-PDT reaction, such as cell killing. Such a marker may provide short-term information about PDT treatment delivery success or failure, information that may be highly useful to the clinician who has to decide whether potential treatment failure requires prompt retreatment of a specific tumor lesion. In this study, we tested the hypothesis that STAT3 cross-linking can act as molecular marker for the local/regional PDT reaction and thus allow early evaluation of the PDT response.
|
Materials and Methods |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionReferences |
|---|
10 days after inoculation when they had reached 6 to 8 mm in diameter. The Roswell Park Cancer Institute Institutional Animal Care and Use Committee approved all procedures carried out in this study. Photosensitizer. Clinical-grade, pyrogen-free 2-[1-hexyloxyethyl]-2-devinyl pyropheophorbide-a (HPPH; ref. 13) was obtained from the Roswell Park Pharmacy and reconstituted to 0.4 mmol/L in pyrogen-free 5% dextrose (D5W; Baxter Corp.) in distilled water containing 2% ethanol and 0.1% Tween 80.
PDT treatment. For in vitro experiments, subconfluent cell monolayers in six-well plates were exposed in the dark to 0 to 2 µmol/L HPPH (exposure conditions indicated in the text). They were then illuminated with 665 nm light at a fluence rate of 14 mW/cm2 for fluences of 1 or 3 J/cm2.
For in vivo experiments, animals were injected via tail vein with 0.4 or 0.8 µmol/kg HPPH. Tumors were exposed to 665 nm light 24 h after HPPH administration. A spot diameter of 1.1 cm covering slightly more than the tumor surface was illuminated. Tumors received graded doses of 665 nm light (10-160 J/cm2) at 10, 14, or 75 mW/cm2. All experiments included control animals treated with either photosensitizer or light alone.
The light source for all PDT treatments consisted of a dye laser (375, Spectra Physics) pumped by an argon ion laser (2080, Spectra Physics) using DCM dye (Exciton). Output from the dye laser was passed through an 8-way beam splitter, and the power of each beam was set individually with Brewster window-type attenuators (14).
Measurement of light attenuation. Light attenuation in 4- to 5-mm-thick MCA205 tumors was measured by in vivo spectroscopy (15). Briefly, the diffuse fluence 
at 665 nm was measured at increasing probe separations using a two-fiber contact probe, one fiber being the light source and the other the receiver. The total attenuation 
is the slope of ln(r) plotted against r, where r is the probe separation in mm and is the diffuse 665 nm fluence escaping from the tumor at r mm from the light source fiber.
Western blot analysis. Cells from in vitro cultures were lysed and tumor tissue was homogenized in radioimmunoprecipitation assay buffer and processed as described (11). Briefly, protein extracts (20-40 µg) were separated on 6% SDS-polyacrylamide gels. On all gels, reference protein markers for loading, molecular size detection, and cross-comparison among gels were included. After separation, proteins were transferred to nitrocellulose membranes (Hybond-P, Amersham Pharmacia Biotech). Nonspecific interactions were blocked by incubating the membranes first for 30 min at room temperature with PBS-0.1% Tween 20-5% skim milk. Membranes were then reacted overnight at 4°C with antibodies to STAT3 (Santa Cruz Biotechnology) and phosphotyrosine-STAT3 (New England Biolabs, Inc.), caspase-3, poly(ADP-ribose) polymerase (Cell Signaling Technology, Inc.), actin (Sigma-Aldrich), or leukemia inhibitory factor receptor (Santa Cruz Biotechnology). Detection of the immune complexes was done using peroxidase-coupled secondary antibodies and enhanced chemiluminescence detection (Pierce Chemical). To quantify the immunodetectable signals, enhanced chemiluminescence images were recorded on X-ray films by various lengths of exposure to ensure recovery of signals that lay in the linear range of detection by digital scanning. The net pixel values of each band were determined by integration using the ImageQuant TL program (Amersham Biosciences). The cross-comparison of separate analyses relied on normalization of the values relative to the coseparated reference markers. Due to the fact that higher-molecular-size STAT3 cross-linked complexes were not always detectable in tumor extracts, the ability of PDT to cross-link STAT3 was expressed only by the percentage conversion of monomeric STAT3 into the dimer (form I on Fig. 1 ). For better visibility of cross-linked proteins, the presentation of data in the figures used images that were exposed longer than those used for quantification.
|
Assessment of tumor response. Following treatment, orthogonal diameters of tumors were measured once every 2 days with calipers. The tumor volume, V, was calculated with the formula V = (lw2 / 2), where l is the longest axis of the tumor and w is the axis perpendicular to l. The tumors were monitored until they reached a volume of >400 mm3, at which time the mice were sacrificed. Regrowing tumors reached the 400 mm3 volume within 10 days. No tumor regrowth was ever observed later than day 50, and therefore, animals were considered cured if they remained tumor-free for at least 60 days after PDT.
Patient samples. Tumor tissue was obtained from patients undergoing bronchoscopic PDT for obstructive endobronchial tumors (non–small cell lung cancer). Treatment and biopsy were approved by the Roswell Park Cancer Institute Institutional Review Board, and informed consent was obtained from all patients. Forty-eight hours before light treatment, the patients had received HPPH (5.3 or 4 mg/m2) by infusion. Light (665 nm) was delivered interstitially at 400 mW/linear cm with a 2-cm diffusing fiber for a total fluence of 130 or 140 J/linear cm. One biopsy was obtained before light exposure, and two biopsies were obtained immediately after completion of treatment; one of these was taken from an area proximal to the light source, the other from an area distal (10 mm, estimated) from the light source and close to the bronchial wall. Samples were immediately processed for Western blotting. Three days after treatment, patients underwent a second bronchoscopy to remove necrotic tissue. At that time, the local PDT response was visually assessed and recorded.
Statistical evaluation. All measured values are presented as mean ± SE. The one-tailed Student's t test was used for comparison between groups in all experiments, except for tumor response determinations. P values of 
0.05 represent statistical significance. For tumor response data analysis, hours to event (i.e., to 400 mm3 tumor volume) was calculated for each animal by linearly interpolating between the times just before and after this volume was reached using log (tumor volume) for the calculations; both tumor volume and hours-to-event calculations were done using Excel (Microsoft). Tumor responses between groups were compared using the Kaplan-Meier analysis. Briefly, the calculated hours-to-event data for individual animals were entered in a Prism (version 3.0; GraphPad Software, Inc.) spreadsheet. Prism calculates and graphs event curves (i.e., the fraction of subjects not reaching the events as a function of time) for each group and calculates the group median time to event. Event curves were compared by the Prism program using the log-rank test, which calculated a two-tailed P value testing the null hypothesis that the curves were identical.
|
Results |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionReferences |
|---|
|
|
15 days at the highest light dose but no long-term tumor control (data not shown).
|
3 mm thickness; Fig. 5
) and adjacent normal tissue from outside the treatment field were harvested immediately following PDT and divided into sections, the position of which with regard to the incident light was recorded (Fig. 5A, a). These samples from the tumor and adjacent dermal layer were extracted and subjected to the Western blotting for the presence of STAT3 cross-links (Fig. 5A, b). Their quantification (Fig. 5A, c) confirmed that the tissue from a position farthest removed from the incident light (i.e., receiving the lowest fluence) contained less cross-links than the tumor samples closest to the light source and that the level was comparable with that of normal tissues harvested from outside but adjacent to the treatment field.
|
160-96 J) is twice that of the bottom portion (Fig. 5B, right lane; fluence range, 96-56 J). To validate the fluence dependence observed in this tumor, another tumor was exposed to an incident fluence of 50 J/cm2 (Fig. 5B, b), the fluence estimated to reach the very bottom of a tumor exposed to 160 J/cm2. Here, the top portion of the tumor (Fig. 5B, b, left lane; fluence range, 50-30J) showed one third of the cross-links observed in the top portion of the tumor exposed to 160 J/cm2, and the extent of cross-linking, as expected, is less than the bottom half of the tumor exposed to 160 J/cm2. In the bottom half of that tumor (Fig. 5B, b, right lane), STAT3 cross-links are barely detectable.
We have reported previously that PDT can not only generate STAT3 cross-links but also lead to the immediate loss of some but not all interleukin-6 family cytokine receptor proteins (7), including the leukemia inhibitory factor receptor-. This additional PDT-dependent event, together with STAT3 cross-linking, became particularly evident in the example of Colon38 tumors (Fig. 5C), which expresses a high basal level of leukemia inhibitory factor receptor-. The spatial distribution (Fig. 5C, a) of STAT3 cross-links coincides with the reduction of leukemia inhibitory factor receptor- (Fig. 5C, c), emphasizing the delivery of a biologically significant PDT dose in tumor areas where STAT3 cross-links are in evidence.
STAT3 cross-links in tumors increase with tumor cures. Having established the relationship of STAT3 cross-linking, PDT dose and, at least in vitro, cell death, we hypothesized that the relative level of STAT3 cross-linking in vivo could provide a measure for the relative cytotoxicity of the PDT treatment and tumor cure. This relationship was tested for MCA205 tumors in which the percentage of conversion of STAT3 to cross-link for individual tumors treated with a range of PDT doses was quantified immediately after treatment (Fig. 6A ). Although cross-linking increased gradually up to 80 J/cm2, it is noteworthy that the only significant increment (P = 0.0086) was between the values for 40 and 80 J/cm2, with no further increase at 120 J/cm2. The tumor response/cure data were obtained from animals that were treated in parallel to those used to determine STAT3 cross-linking (Fig. 6B). In agreement with the cross-linking data, a significant increase in long-term tumor control was observed between 40 and 80 J/cm2 (P < 0.0001), with no significant further improvement at 120 J/cm2.
|
|
|
Discussion |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionReferences |
|---|
It is widely recognized that limitations to tumor cell kill by PDT exist in vivo, the most important of which being (a) inhomogeneous photosensitizer distribution within and among tumors (17), including a gradual decrease of photosensitizer concentration with distance from blood vessels (18); (b) insufficient light penetration through the tissue (16, 19); and (c) insufficient oxygen availability (20, 21). Although preclinical studies have successfully defined many of these limitations, showing the frequent regional heterogeneity of the PDT effect, detailed information about these issues in human tumors undergoing PDT is extremely scant. Several studies have searched for the presence of predictive PDT response markers. Koukourakis et al. (22) in a retrospective study of archival material from 37 patients with esophageal cancer reported that Bcl-2 expression was associated with a high complete response rate. The authors suggested that the reason for this association was the earlier observation by others that Bcl-2 is a target for PDT destruction (23, 24). On the other hand, the expression in the same material of hypoxia-inducible factor 1 and 2, which reflects the level of tumor hypoxia, was associated with poor PDT response. The latter association is to be expected because PDT critically depends on the presence of oxygen (25). However, even if it were proven that the degree of expression of these molecules was predictive of ultimate PDT outcome, these assessments would not furnish information about the actual short-term PDT response.
These complexities taken together make it unlikely that expression of a "predictive" factor determined before PDT would be sufficient for early evaluation of PDT response. Such early evaluation, however, is a clinical need as it may inform whether a tissue has received an effective PDT exposure or whether it may need further treatment to achieve tumor control.
We have recently made the discovery that cross-linking of STAT3 showed extraordinary PDT dose dependence (11). We now present preclinical and clinical evidence that STAT3 cross-links may serve as a "dosimeter" for absorbed PDT dose and a valuable biomarker to give short-term information about the local/regional PDT response. In vitro, strong correlations were observed between cell kill and STAT3 cross-linking, with a tendency for increases in cross-links to diminish at very high PDT doses. The correlation was strong whether HPPH exposure of cells was brief or prolonged and whether a clonogenic or 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide assay was used as survival end point, attesting to the generality of the cross-linking reaction. The upper limits in detecting STAT3 conversion at high PDT doses may be determined in part by secondary cross-linking reactions that shift immune detectable STAT3 into the high-molecular-size region of the gel (Fig. 1) and the concurrent initiation of tissue degradation with loss of protein integrity due to proteolysis. Although correlations between cell kill and STAT3 cross-linking were high, no relative cross-link values could be assigned to specific levels of cell kill across the panel of cell lines tested. However, a STAT3 conversion of >25% generally related to 100% cell kill. Furthermore, the appearance of strong bands of cross-linked STAT3 coincided with the appearance of cleaved poly(ADP-ribose) polymerase and apoptosis. PDT dose-dependent STAT3 cross-linking occurred across the entire spectrum of cells tested, including cell types analyzed earlier, such as FaDu, a human hypopharyngeal carcinoma–derived cell line (11). Only the human prostate cell line PC3, which lacks STAT3, did not show protein forms that reacted with anti-STAT3.
The experiments carried out in preclinical tumor models reveal two important aspects, namely the absence of any STAT3 cross-linking in photosensitizer-containing tumor tissue in the absence of light exposure (Fig. 4, contralateral tumors) and the relatively high reproducibility of results among different animals. Examination of tumor subsections suggests the expected relationship between STAT3 cross-linking and light fluence (i.e., distance from the light source). As fluence of light penetrating through tissue decreases exponentially as a function of distance from the light source (16, 19), cross-linking was least at the bottom of tumors illuminated from the top. Cross-links were also detectable, although at lower levels, in the normal tissue adjacent to the treatment field, indicating that stray light diffusing through the tissue was sufficient to cause a biological effect and also showing the sensitivity of STAT3 cross-links in reporting low PDT exposures. Although there are numerous variables that can limit the PDT reaction, it is clear that in these small and relatively homogeneous experimental tumors light dose is the dominant factor determining tissue response.
PDT has been shown in vitro to cause the loss/degradation of membrane receptor proteins, such as growth factor (26) and cytokine receptors (7). We now show that in regions of strong STAT3 cross-linking in vivo cytokine signaling can be compromised not only by the reduction of available STAT3 but also by the reduction/elimination of membrane cytokine receptors. Thus, strong STAT3 cross-linking indicates the delivery of a PDT dose sufficient to initiate other biological events, such as apoptosis and changes in signal transduction.
Side by side comparisons of STAT3 cross-linking in PDT-treated tumors and effectiveness of identical treatments in parallel experiments of tumor control also show a strong relationship between STAT3 conversion and tumor cure, with significant increases in cross-linked STAT3 corresponding to a significant increase in tumor cures. The highest percentage of cross-linked STAT3 in tumors (highest value, 25%; median, 12%) was less than observed in cell monolayers. This can be explained by the fact that the measured values are the average of the whole tumor, including regions that were distant from the light source and received less PDT dose.
The feasibility of using STAT3 cross-links as a "dosimeter" for the PDT response was evaluated in five of six patients undergoing PDT for endobronchial tumors. As expected, the percentage of conversion from monomeric STAT3 to STAT3 complexes in most cases decreased with distance from the light source, reflecting the fall-off in fluence. The variability in the values for STAT3 cross-links, however, was large. This may reflect differences in the light penetrance through the tumor due to bleeding or other heterogeneities, such as in tissue oxygenation. It also may in part simply be due to the lack of precision in tumor sampling inherent in the procedure. However, most importantly, the negative results obtained from patient 6, the only patient who did not respond to treatment, indicate that lack of STAT3 cross-links and other proteins detectable immediately after PDT treatment correctly predicted treatment failure. In our clinical experience with the treatment of a wide range of solid tumors, complete lack of response is seen rarely. It is thus remarkable that this nonresponse was reflected in the STAT3 cross-link assay. With photosensitizers that have relatively slow pharmacokinetics, such as Photofrin and, to a lesser degree, HPPH, it is possible to retreat without drug reinjection if regional or complete treatment failure is identified early. The described assay might help in this early detection of treatment failure. This approach also might permit evaluation of procedure-associated photoreactions, such as illumination of the treatment site with inspecting light sources causing appearance of cross-linked STAT3 in "pretreatment" samples as well as of unintended light exposure of normal tissue.
In summary, the results from our in vitro, in vivo, and preliminary clinical studies strongly support the value of STAT3 cross-linking as measure for PDT reaction. They further provide evidence for the potential value of STAT3 cross-links as a molecular marker capable of early prediction of treatment success or failure. Accurate light dosimetry is still a major challenge in the clinical application of PDT, especially when applied in hollow organs with poor symmetry. Although systems are being developed that can directly measure in situ the cytotoxic agent (singlet oxygen) generated by a given treatment, these are not yet clinically available. Therefore, the identification of a surrogate molecular marker for absorbed PDT dose is of high clinical relevance.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Note: B.W. Henderson and H. Baumann contributed equally to this work.
Current address for C. Daroqui: Department of Oncology, Albert Einstein Cancer Center, Montefiore Medical Center, Bronx, NY 10467.
Received 12/13/06; revised 2/27/07; accepted 3/ 8/07.
|
References |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionReferences |
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Cancer Research | Clinical Cancer Research |
| Cancer Epidemiology Biomarkers & Prevention | Molecular Cancer Therapeutics |
| Molecular Cancer Research | Cancer Prevention Research |
| Cancer Prevention Journals Portal | Cancer Reviews Online |
| Annual Meeting Education Book | Meeting Abstracts Online |
) or 3 J/cm2 (
). Identically treated cultures were divided into two groups: one group was used for cell survival assay (left) and the other group for STAT3 immunoblotting (right), respectively. MCA205 cell survival was determined by clonogenic assay, and Colon38 and A549 cell survival was determined by 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide assay. Data represent the results from at least two independent experiments.
