Influence of Oct1/Oct2-Deficiency on Cisplatin-Induced Changes in Urinary N-Acetyl-β-D-Glucosaminidase

This article is featured in Highlights of This Issue, p. 4067

Cisplatin is a DNA-crosslinking agent that is among the most widely used anticancer drugs worldwide in both adult and pediatric populations. The dose-limiting side effects associated with cisplatin-based chemotherapy include renal tubular dysfunction (nephrotoxicity), and to a lesser extent, hearing loss (ototoxicity; ref. 1). Severe and irreversible damage to the kidney is a tremendous health problem and remains the single most important complication of cisplatin treatment as it may limit further treatment or even threaten life. This side effect primarily affects the S3 segment of the renal proximal tubules and occurs in up to 40% of patients despite intensive prophylactic measures (2), including extensive prehydration and posthydration regimens with hypertonic saline.

Over the last three decades, various approaches have been reported to afford renoprotection during cisplatin treatment, although most of these have not been evaluated in animal models or humans (3). The clinical application of many of these strategies has been hampered by the recognition that (a) cisplatin has multiple intracellular targets and hence blocking a single injurious event will only have partial protective effects in the kidney, and (b) the renoprotective approach may diminish the antitumor effects of cisplatin given the overlap in cell death signaling pathways between normal kidney cells and tumor cells (3). Therefore, an ideal approach of renoprotection is to protect the kidneys without affecting the therapeutic effects in tumors.

Using transfected mammalian cells, we previously reported that cisplatin is a substrate for the organic cation transporter OCT2 (4), which is highly expressed at the basolateral membrane of renal tubular epithelial cells and is considered a major transporter in the active secretion of xenobiotics from the circulation into the kidney. More recently, we found that mice lacking the ortholog transporters Oct1 and Oct2 [Oct1/2(-/-) mice] have impaired urinary excretion of cisplatin and that these animals are protected from experiencing severe cisplatin nephrotoxicity (5). This suggests that these transporters are key proteins involved in the initial step toward renal elimination of cisplatin. Based on these observations, we hypothesized that the extent to which cisplatin interacts with Oct 1 and Oct2 (or OCT2 in humans) and subsequently affects nephrotoxicity is dependent on intracellular drug concentrations in the renal tubule and subsequent activation of pathways resulting in tubular apoptosis. The objectives of the current study were to (a) assess the extent of renal secretion of cisplatin in wild-type mice and Oct1/2(-/-) mice, (b) determine the influence of cisplatin treatment on the time course of biomarkers of nephrotoxicity and overall survival, and (c) evaluate the effects of OCT2 inhibitors on cisplatin transport and nephrotoxicity.

The Flp-In transfection system, DMEM, PBS, Lipofectamine 2000, hygromycin, zeocin, Opti-MEM reduced serum medium, TRIzol, Superscript III first strand synthesis system, and fetal bovine serum were all obtained from Invitrogen. The human full length cDNA clone of OCT2 was purchased from Origene. Cisplatin and [¹⁴C]-tetraethylammonium bromide (TEA) were obtained from Sigma Chemical Co. and American Radiolabeled Chemicals, respectively. All other compounds used in this study were of reagent grade or better.

Adult (8-12 weeks old) male FVB wild-type and Oct1/2(-/-) mice were purchased from Taconic. All animals were housed and handled in accordance with the Institutional Animal Care and Use Committee of St. Jude Children's Research Hospital. Animals were housed in a temperature-controlled environment with a 12-hour light cycle and were given a standard diet and water ad libitum.

Blood samples were obtained at 0.25, 0.75, 1, 2, 4, and 8 hours after an i.p. dose of cisplatin at 10 mg/kg, centrifuged to obtain plasma, and flameless atomic absorption spectrometry (AAS) was used to determine total platinum concentrations (5). AUC_(0-24) was calculated using noncompartmental pharmacokinetic analysis extrapolated out to 24 hours. The unbound platinum concentration was determined following ethanolic plasma-protein precipitation (6). Pharmacokinetic parameters were calculated using the software package PK Solutions 2.0 (Summit Research Services).

For experiments involving collection of urine, Oct1/2(-/-) and wild-type (n = 12/group) animals were single-housed in metabolic cages in a temperature-controlled environment with a 12-hour reverse light cycle. After the animals had acclimated to metabolic cages, a 24-hour baseline urine sample was collected 16 hours after the onset of light. The next day, the mice were given a single i.p. injection of cisplatin at a dose of 10 mg/kg. Urine was collected before drug administration and at 4, 24, and 48 hours after administration. Urine samples were diluted with nitric acid (0.2%) and analyzed for total platinum using AAS (5). An estimate of the renal clearance (CL_R) of total platinum was determined as follows: CL_R = A_{urine (0-24)}/AUC_(0-24), where A_{urine (0-24)} and AUC_(0-24) are the cumulative urinary excretion of platinum up to 24 hours and the area under the plasma concentration-time curve for protein-unbound platinum up to 24 hours, calculated using the linear trapezoidal rule, respectively. Metabolic cages were washed with a double wash cycle in a portable dishwasher using hot water and household dishwashing detergent between experiments done in different animals.

Glomerular filtration rate (GFR) at baseline was estimated in male FVB mice and Oct1/2(-/-) mice from body weight (BW) from intraspecies allometry using the following equation: GFR = 0.036 × BW^{0.74 ± 0.15} (7). The validity of this approach is supported by the previous finding that GFR determined from [¹⁴C]inulin clearance was not significantly different between wild-type mice and Oct1/2(-/-) mice, and corresponded well to the GFR estimated from BW alone (8). Levels of blood urea nitrogen (BUN), serum creatinine, and urinary creatinine were determined using a Vetscan autoanalyzer (Abaxis). Urinary NAG activity was chosen due to it being an effective and sensitive biomarker for nephrotoxicity (9). Urinary NAG activity was measured before and after cisplatin treatment was measured in wild-type and Oct1/2(-/-) mice using a kit from Diazyme Laboratories (Poway) with detection at 505 nm on a μQuant microplate spectrophotometer (BioTek Instruments). Subsequently cisplatin-induced (10 mg/kg, i.p.) urinary NAG activity was determined in wild-type mice administered a saline vehicle control (i.v.), cimetidine (30 mg/kg, i.v.), or nothing.

Overall survival was assessed in male wild-type mice and age- and weight-matched male Oct1/2(-/-) mice after administration of cisplatin by i.p. injection at a single dose of 10 or 20 mg/kg. Overall survival was assessed by the Kaplan-Meier method, using an exact log-rank test and a Cox-Mantel proportional hazard model to determine the impact of mouse genotype and nephrotoxicity.

For microscopic examination, tissues were fixed in 4% phosphate-buffered formalin, embedded in paraffin, sectioned at 4 μm, and stained with H&E in accordance with standard procedures.

The Flp-In transfection system was used as described to produce stably transfected 293Flp-In cells expressing OCT2 or an empty pcDNA5 vector (4). These cell lines were collected and reseeded in complete DMEM containing 10% fetal bovine serum and hygromycin B (100μg/mL). All uptake experiments were done on monolayer cultures in 6-well plates at 37°C, as described (4). For the transport inhibition assays, monolayers were washed once with warm PBS then preincubated with DMEM containing potential OCT2 inhibitors at initial concentrations of 1 mmol/L. This medium was removed and replaced with the same concentration of inhibitor in DMEM solution with [¹⁴C]-TEA (2 μmol/L) and allowed to incubate for an additional 30 minutes at 37°C. After incubation, cells were washed twice with cold PBS, collected and solubilized in 1N sodium hydroxide. Radioactivity was assessed by liquid scintillation counting, platinum levels were measured using AAS (5), and protein concentrations were measured using a bicinchoninic acid (BCA) protein-assay kit (Pierce Biotechnology). Compounds that were found to be significant inhibitors of TEA uptake were subsequently tested at a concentration of 1 μmol/L.

The potential functional change in transport associated with a common OCT2 variant (p.270Ala>Ser) was assessed using cisplatin transport in the presence and absence of cimetidine. Briefly, HEK293 cells expressing either an empty vector (pcDNA3.1), wild-type OCT2 (containing reference sequence), or a protein variant of OCT2 containing the p.270Ala>Ser amino acid substitution were maintained and tested under similar conditions mentioned above. These cells are stably transfected and have been previously reported on (10).

RNA from the NCI₆₀ cancer cell lines were provided by the National Cancer Institute tumor repository (Bethesda). SK-OV-3 cells were from the American Type Culture Collection. RNA was reverse transcribed using SuperScript III first strand synthesis supermix for quantitative reverse transcriptase-PCR (Invitrogen) according to the manufacturer's recommendations. Gene transcripts were quantified using SYBR Green PCR mastermix (Qiagen) and primers previously described (11). Reactions were carried out in triplicate unless otherwise stated as previously reported (4). Briefly, 25 μL volumes were used with the following PCR variables: 95°C for 15 minutes then 40 cycles of 94°C for 30 seconds, 60°C for 30 seconds, and 72°C for 30 seconds followed by dissociation cycles. Transcripts of each sample were normalized to the housekeeping gene, GAPDH.

Primers for SLC22A2 were designed as previously described (4). Samples were cleaned with ExoSAP-IT reagent (USB Corporation) following PCR and sequenced in both forward and reverse directions using Big Dye Terminator version 3.1 on an Applied Biosystems 3730XL DNA analyzer. Sequencher software version 4.7 (Gene Codes Corporation) was used for sequencing analysis.

All data are presented as mean and SE, unless otherwise stated. Statistical analyses were done using either ANOVA, followed by Bonferroni posthoc test, or a two-tailed t test where appropriate, and P values < 0.05 were considered to be statistically significant. All statistical calculations were done using the software package NCSS version 2004 (Number Cruncher Statistical System).

It was recently reported that in the mouse, cisplatin is recognized as a substrate by both Oct1 and Oct2 (4, 5, 12–14). To further investigate the roles of Oct1 and Oct2 in renal secretion of cisplatin, we compared the plasma and urinary pharmacokinetics of this drug in Oct1/2(-/-) mice. The renal clearance of platinum was greatly reduced in the Oct1/2(-/-) mice (15.8 ± 2.16 versus 25.9 ± 2.24 mL/h in wild-type; P < 0.01), although there were no differences in the estimated GFR at baseline (16.4 ± 0.21 versus 16.8 ± 0.52 mL/h). The ratio of renal clearance to GFR was about 1.54 ± 0.13 for wild-type mice, which is comparable with what has been found by others (15). In Oct1/2(-/-) mice, however, this ratio was reduced to 0.96 ± 0.13, indicating that the net tubular secretion of platinum was completely abolished in these animals (P < 0.01).

We found that in wild-type mice receiving cisplatin, the widely used biomarkers for assessing cisplatin nephrotoxicity, BUN and serum creatinine, are less than ideal because increases only occur after substantial kidney damage, and then with a time delay (Fig. 1A and B), consistent with previous findings (16). Indeed, BUN and serum creatinine did not show significant elevation in the mice until 72 hours after administration of cisplatin. This is despite the notion that histopathologic analysis indicated proximal tubular damage as early as 24 hours following drug administration (Fig. 1A, right, and B).

As an alternative to BUN and creatinine, we explored the utility of measuring urinary NAG for evaluating the role of renal transporters in cisplatin nephrotoxicity. In contrast to BUN and creatinine, we found that NAG is a sensitive quantitative biomarker for detection of cisplatin nephrotoxicity with significant changes in activity occurring as early as 4 hours (P < 0.0001; Fig. 2A). Cisplatin-induced changes in urinary NAG activity were much less pronounced in Oct1/2(-/-) mice compared with wild-type mice over the studied time interval (Fig. 2A), consistent with the reduced nephrotoxicity observed in the transporter-deficient animals (5). Additional analysis on mice following a single dose of cisplatin indicated that a cutoff point of >0.4 AU for the cumulative urinary NAG activity over 48 hours is highly predictive of severe nephrotoxicity, with a “severe/not-severe” odds ratio of 21.0 (95% confidence interval, 2.91-151; P = 0.0017).

The absence of severe cisplatin nephrotoxicity (i.e., when >75% of tubules are damaged) in Oct1/2(-/-) mice is a critical difference from the observed side effects in wild-type mice, because the severity of renal damage is linked with mortality following a single dose of cisplatin. Indeed, the decreased toxicity in the Oct1/2(-/-) mice is associated with a significantly prolonged overall survival (Fig. 2B). For example, at a cisplatin dose of 10 mg/kg, the wild-type/Oct1/2(-/-) Cox-Mantel hazard ratio was found to be 8.1 (95% confidence interval, 2.1-31; P = 0.0078). Interestingly, although all wild-type mice died by day 7, a subset of Oct1/2(-/-) mice seemed to be protected from cisplatin-induced death, presumably due to reduced cisplatin-induced nephrotoxicity.

To explore the possible utility of chemical inhibitors of OCT2 as a strategy to prevent severe cisplatin nephrotoxicity, we screened 16 agents at varying concentrations for OCT2 inhibitory potential in transfected 293Flp-In cells using TEA as a test substrate. As shown in Supplementary Fig. S1, at a substrate to inhibitor concentration ratio of 1:500, all of the known OCT2 inhibitors (amantadine, cimetidine, diclofenac, imatinib, and nicotine) used in this screen showed significant inhibition of OCT2 function, as determined by intracellular accumulation of [¹⁴C]-TEA (8-42% of control values). A number of supportive care drugs (aprepitant, granisetron, metoclopramide, omeprazole, and pantoprazole) and chemotherapeutic agents (docetaxel and etoposide) also showed significant inhibition of OCT2 function (by 1-29% of control). When tested at clinically relevant concentrations (1 μmol/L), however, only imatinib (75% of control, P = 0.012) and cimetidine (70% of control, P = 0.0006) were found to be inhibitors of OCT2 function (Fig. 3).

The ability of cimetidine to inhibit OCT2 function was further confirmed using cisplatin as a test substrate in three different OCT2-overexpressing cell lines (Fig. 4). These results indicated that cimetidine, at a substrate to inhibitor concentration ratio of 1:2, significantly reduced the uptake of cisplatin almost 4-fold (P < 0.0001). Interestingly, cisplatin transport in the absence or presence of cimetidine was not altered in HEK293 cells transfected with a common OCT2 protein variant (p.270Ala>Ser) associated with the rs316019 (808G>T) single nucleotide polymorphism (Fig. 4). This suggests that this particular OCT2 variant is not associated with altered substrate recognition for cisplatin.

To obtain preliminary evidence for the usefulness of adding cimetidine to cisplatin-based therapy to ameliorate nephrotoxicity, urinary NAG activity was determined in mice receiving cisplatin (10 mg/kg i.p.) in the presence and absence of cimetidine (30 mg/kg i.v.; Fig. 5). There was a trend for slightly reduced NAG activity in wild-type mice in which cisplatin administration was preceded by an i.v. injection of saline. Although this difference was not statistically significant (P = 0.19), the observation is consistent with prior clinical data suggesting that i.v. hydration can reduce cisplatin nephrotoxicity to some extent (17). Most importantly, wild-type mice receiving concurrent cimetidine with cisplatin had urinary NAG values that were significantly reduced compared with mice receiving cisplatin only (P = 0.016) and were similar to levels in cisplatin treated Oct1/2(-/-) mice, as shown in Fig. 2A (0.2697 ± 0.082 and 0.2624 ± 0.141, respectively; P = 0.91). This suggests that cimetidine is able to completely inhibit Oct1/Oct2-mediated uptake of cisplatin in renal proximal tubular cells, and subsequently ameliorate cisplatin nephrotoxicity.

The contribution of OCT2 to the uptake of cisplatin in human tumor cells is currently unknown. We previously reported that the OCT2 gene SLC22A2 was either absent or detectable at very low levels in multiple human tumors, including those for which cisplatin is clinically used, such as lung cancer and ovarian cancer (5). In the NCI₆₀ cancer cell line panel, we identified the ovarian cancer cell line SK-OV-3 as having the highest expression of SLC22A2 (Supplementary Fig. S2). However, we found that cellular sensitivity to cisplatin in the NCI₆₀ cancer cell line panel was not significantly associated with the expression of SLC22A2 (R² = 0.009, P = 0.47; data not shown). We also found that expression of SLC22A2 was approximately 175-fold lower in SK-OV-3 cells compared with our OCT2-transfected 293Flp-In cells, and that the expression of other genes of putative relevance to cisplatin transport, such as SLC22A1 (encoding OCT1), were very low in all of the cell models tested (Fig. 6A). Although the absolute uptake of cisplatin in SK-OV-3 cells was quite substantial, the presence of an excess amount of cimetidine had no influence on the cellular uptake and retention of cisplatin in this model (Fig. 6B). This finding is consistent with the possibility that substantial overexpression of OCT2, such as that observed in our transfected 293Flp-In cells, or under normal physiologic conditions in human kidney, is required before its quantitative contribution to cisplatin transport can be discerned.

This study directly shows in vivo that organic cation transporters (OCT2 in humans, Oct1 and Oct2 in mice) are essential for the active secretion of cisplatin into renal proximal tubular cells, and that these proteins play a crucial role in the development of cisplatin nephrotoxicity. Our collective in vitro and in vivo data have potentially important clinical implications for the optimization of cisplatin usage and strongly support the hypothesis that pharmacologic inhibitors of OCT2 can be used to prevent cisplatin-induced kidney damage.

A role of organic cation transporters in the renal uptake of cisplatin has been suggested previously by several studies. This was initially deduced from in vitro studies that showed that cisplatin could inhibit the cellular uptake of the prototypical OCT2 substrate TEA (18, 19). Furthermore, in vivo administration of cisplatin inhibits the renal clearance of other organic cations (20), and conversely, several organic cations can affect the renal clearance of cisplatin (21) without affecting GFR (22). Recently, in vitro methods using OCT2-overexpressing cells have been used to confirm that cisplatin is itself a substrate for this transporter (4). The role of Oct2 in cisplatin-induced nephrotoxicity was suggested by the increased nephrotoxicity in male rats, which have higher expression of Oct2 in the kidneys, as compared with females (23). Moreover, our previous study indicates that functional loss of the mouse ortholog transporters Oct1 and Oct2 is associated with a marked reduction in urinary excretion of platinum and nephrotoxicity, as determined by histopathologic staining (5). This renal protection with the deletion of Oct1 and Oct2 has recently been confirmed in a study evaluating cisplatin-induced toxicities (24). Our current findings that Oct1/2(-/-) mice lack renal secretion of platinum and also have an increase in the median survival time compared with wild-type mice receiving the same cisplatin dose further highlight the importance of Oct1 and Oct2 in the pharmacology and toxicology of cisplatin.

It has been long recognized that better biomarkers for cisplatin nephrotoxicity are needed both for animal studies and for use in humans where early detection of kidney injury will influence therapy and potentially morbidity and mortality. A recent report on the effects of cimetidine as an inhibitor of cisplatin-induced ototoxicity and nephrotoxicity also found inconsistencies between differing biomarkers of nephrotoxicity (24). Cimetidine pretreatment reduced levels of several biomarkers for nephrotoxicity when evaluated 4 days after administration (i.e., BUN, tubular protein casts, vacuolization), but this was not seen in all biomarkers studied (terminal deoxynucleotidyl transferase–mediated dUTP nick end labeling staining and proteinurea). This supports the need for a sensitive biomarker that can detect nephrotoxicity early. Here, we found that routinely used measures of kidney injury, such as BUN and serum creatinine, increase significantly only after substantial damage occurs and then with a substantial time delay, which renders BUN and creatinine useless as biomarkers in the context of our studies.

Beyond serum markers, many urinary proteins have been previously evaluated as putative noninvasive indicators of general renal injury, including α- and π-glutathione-S-transferases, neutrophil gelatinase-associated lipocalin, cysteine-rich protein 61, interleukin-18, clusterin, and F-actin (25). However, problems with the reliability of these proteins to identify and monitor kidney injury include instability in the urine, modification due to physicochemical properties of the urine, absence of sustained elevation, and lack of a high-throughput detection method. Based on these considerations, we evaluated cisplatin-induced changes in urinary NAG, a high-molecular-weight lysosomal enzyme that is abundant in the proximal tubular epithelia (26, 27). Due to its large size this molecule is not known to undergo filtration, but rather is released into the urine when renal proximal tubular cells are damaged (26). Urinary NAG activity has been previously used in evaluating lead-mediated damage (27) and cisplatin nephrotoxicity in rats (23). These prior studies and our current work in wild-type and Oct1/2(-/-) mice strongly support the hypothesis that urinary NAG activity is a highly sensitive, stable, and early biomarker for cisplatin nephrotoxicity, especially compared with BUN and creatinine. The ability of cimetidine to block Oct1/Oct2-mediated uptake of cisplatin in renal proximal tubular cells, and subsequently ameliorate cisplatin-induced changes in NAG activity, support further efforts to explore the use of this biomarker in rodents and patients receiving cisplatin.

Previously, cimetidine was found in vitro to decrease the cytotoxic effects of another platinum-based chemotherapy, oxaliplatin, in cells transfected to express OCT1, OCT2, or OCT3 (14). It should be noted, however, that although this study did find increased uptake of cisplatin in cells overexpressing OCT1 or OCT2, the effect was not as robust as in other studies (4, 13). This is most likely due to differences in transfection models and efficiency, as suggested by the authors. Although cisplatin was not as robust a substrate for these transporters as oxaliplatin, there was still increased sensitivity to cisplatin-induced toxicity in OCT1- and OCT2-overexpressing cells, further supporting the hypothesis that inhibiting OCT-mediated uptake of platinum-based chemotherapeutics can ameliorate the toxicity of these agents. Although comprehensive studies comparing mouse and human pharmacokinetics of cimetidine are not available, it should be noted that using the Food and Drug Administration (FDA) guidelines for determining human equivalent doses, a 30 mg/kg dose of cimetidine used in this study is comparable to a 200 mg dose in a human with a body surface area of ∼2 m² (FDA. Guidance for determining human equivalent doses. http://wwwfdagov/downloads/Drugs/GuidanceComplianceRegulatoryInformation/Guidances/ucm078932pdf.), which is based on interspecies conversion factors previously established (28).

Most of the approaches reported in the literature to afford renal protection during cisplatin treatment have not separately considered the possible implications on the anticancer actions of cisplatin in tumors. It is therefore imperative to conclusively show that intentional inhibition of OCT2 function does not negatively influence cisplatin effects on the tumor. The mechanism of cellular uptake of cisplatin in tumor cells is not entirely clear and may vary from one cell type to another (3). Preclinical studies have shown that a high-affinity transporter known as copper transporter 1 (CTR1, SLC31A1) plays a critical role in the uptake of cisplatin in yeast and several tumor cell lines. However, the accumulation of cisplatin in human cancer cells expressing this transporter is limited by the fact that cisplatin triggers the downregulation and proteasomal degradation of CTR1, thereby limiting its own uptake (29). Our current studies show that even the presence of an excess amount of cimetidine had no influence on the cellular uptake and retention of cisplatin in SK-OV-3 cancer cells, which carry the highest SLC22A2 expression among the NCI₆₀ cell lines. In addition to this in vitro observation, it is noteworthy that a quantitative real-time PCR analysis of 80 ovarian tumor specimens revealed recently that the vast majority of tumors had low (n = 12) or undetectable (n = 68) levels of SLC22A2 (30). Of the 80 specimens, 51 were obtained from patients treated with cisplatin-containing chemotherapy, 42 of whom responded to the treatment (30). No difference, either in expression level or frequency of SLC22A2-positive tumors, was apparent between the nonresponders and the responders. These in vitro and ex vivo results suggest that OCT2 function is not a critical determinant of the antitumor effects of cisplatin, and thus that the use of OCT2 inhibitors should not affect the intracellular accumulation of cisplatin into tumors.

Collectively, our study suggests that the use of a pharmacologic inhibitor of OCT2 is a feasible mechanism to exploit for preventing cisplatin nephrotoxicity without influencing effects on the tumor. Furthermore, this study indicates that urinary NAG activity serves as a sensitive early biomarker for nephrotoxicity. If this dose-limiting toxicity can be alleviated in patients, the therapeutic window of cisplatin in cancer treatment can be dramatically improved. We are currently conducting an ongoing study to further elucidate the role of cimetidine in long-term survival, tumor suppression in xenograft models, and a prospective clinical trial to evaluate the effects of cimetidine as a conjunctive therapy to reduce or eliminate cisplatin nephrotoxicity. Although cimetidine will be used here as a pharmacologic tool for establishing proof of concept, our future studies will be aimed at developing and evaluating more selective agents that could provide better protection against the renal damage associated with cisplatin use.

No potential conflicts of interest were disclosed.

We thank John Killmar, Peter Vogel, and Laura Janke for assistance with data analysis, and the National Cancer Institute Tumor Repository for providing RNA and DNA from the NCI₆₀ cell lines.

Grant Support: American Lebanese Syrian Associated Charities (ALSAC) and United States Public Health Service Cancer Center Support Grant 3P30CA021765.

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.