Abstract P3-07-49: Residual cancer burden (RCB) with veliparib/carboplatin in the I-SPY2 trial

Background: I-SPY2 is a multicenter phase 2 trial in high risk stage II/III breast cancer (BC) using adaptive randomization within biomarker subtypes to evaluate novel agents added to standard neoadjuvant chemotherapy. The first regimen to graduate based on the predicted probability of a higher pCR rate within predefined subsets was veliparib/carboplatin + paclitaxel (VC+T→AC vs T→AC) in triple negative BC (TNBC). In TNBC the residual cancer burden (RCB) is prognostic, whether as a continuous index or grouped into classes, with pCR (RCB-0) and RCB-I classes having identical survival. Therefore, we evaluated the use of RCB to further discriminate between investigational and control arms.

Methods: Site pathologists reported RCB for 99% of subjects in the primary efficacy analysis based on pCR (n=114/115). We compared the distribution of RCB reported as a continuous index in each treatment-subset combination to matched concurrently randomized controls using the Wilcoxon rank sum test for RCB index, and Fisher's Exact test for RCB classes (RCB-0/I vs RCB-II/III). The statistics are descriptive rather than inferential, and given the small sample size have no claim on generalizability. We modified the Bayesian model used to compute the estimated probability of success in a future, randomized, phase 3 trial of 300 subjects, if response were defined by either pCR or RCB-I (RCB0/I), or separately if it were defined by pCR alone.

Results: VC+T→AC led to a significantly lower RCB index than T→AC in TNBC (p=0.0021), with a near-significant trend when those with pCR were excluded (p=0.06). There was no significant difference in RCB distributions in the other breast cancer subtypes treated. In TNBC, the odds ratio (OR) for achieving RCB-0/I in the VC+T→AC arm vs control was 8.2 (95% confidence interval (CI): 2.1–35), whereas the OR for achieving pCR was 4.56 (95% CI: 1.25–19.53). The simulations using response information from I-SPY2 to predict the probability of success for VC+T→AC for TNBC in a future phase 3 trial estimated this probability to be 0.99 if modeled using RCB-0/I as the response endpoint, and 0.90 if modeled using pCR as the response endpoint.

Conclusions: Use of RCB index and classes provided additional insight into the effect of adding VC to T, appearing to magnify the improved treatment response that had been observed with pCR rates in TNBC. It will be important to test in randomized trials whether a decrease in the RCB index relative to controls, and/or increased rates of RCB-0/I class, are predictive of survival benefit in TNBC.

Citation Format: Liu MC, Symmans WF, Yau C, Chen Y-Y, Rugo HS, Olopade OF, Datnow B, Chen B, Feldman M, Kallakury B, Hasteh F, Tickman R, Ritter J, Troxel M, Mhawech-Fauceglia P, Duan X, Berry D, Esserman L, DeMichele A. Residual cancer burden (RCB) with veliparib/carboplatin in the I-SPY2 trial. [abstract]. In: Proceedings of the Thirty-Eighth Annual CTRC-AACR San Antonio Breast Cancer Symposium: 2015 Dec 8-12; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2016;76(4 Suppl):Abstract nr P3-07-49.

Inhibition of the Aurora A kinase augments the anti-tumor efficacy of oncolytic measles virotherapy

Oncolytic measles virus (MV) strains have demonstrated broad spectrum preclinical anti-tumor, including breast cancer. Aurora A kinase controls mitotic spindle formation and plays a critical role in malignant transformation. We hypothesized that, by causing mitotic arrest, the Aurora A kinase inhibitor MLN8237 (alisertib) can increase MV oncolytic effect and efficacy. Alisertib enhanced MV oncolysis in vitro and significantly improved outcome in vivo against breast cancer xenografts. In a disseminated MDA-231-lu-P4 lung metastatic model, the MV/alisertib combination treatment markedly increased median survival to 82.5 days with 20% of the animals being long term survivors vs. 48 days median survival for the control animals. Similarly, in a pleural effusion model of advanced breast cancer, the MV/alisertib combination significantly improved outcome with a 74.5 day median survival versus the single agent groups (57 and 40 days respectively). Increased viral gene expression and IL-24 upregulation were demonstrated, representing possible mechanisms for the observed increase in antitumor effect. Inhibiting Aurora A kinase with alisertib represents a novel approach to enhance measles virus-mediated oncolysis and antitumor effect. Both oncolytic MV strains and alisertib are currently tested in clinical trials, this study therefore provides the basis for translational applications of this combinatorial strategy in the treatment of patients with advanced breast cancer.

Predicting response and survival in chemotherapy-treated triple-negative breast cancer

BACKGROUND

In this study, we evaluated the ability of gene expression profiles to predict chemotherapy response and survival in triple-negative breast cancer (TNBC).

METHODS

Gene expression and clinical-pathological data were evaluated in five independent cohorts, including three randomised clinical trials for a total of 1055 patients with TNBC, basal-like disease (BLBC) or both. Previously defined intrinsic molecular subtype and a proliferation signature were determined and tested. Each signature was tested using multivariable logistic regression models (for pCR (pathological complete response)) and Cox models (for survival). Within TNBC, interactions between each signature and the basal-like subtype (vs other subtypes) for predicting either pCR or survival were investigated.

RESULTS

Within TNBC, all intrinsic subtypes were identified but BLBC predominated (55-81%). Significant associations between genomic signatures and response and survival after chemotherapy were only identified within BLBC and not within TNBC as a whole. In particular, high expression of a previously identified proliferation signature, or low expression of the luminal A signature, was found independently associated with pCR and improved survival following chemotherapy across different cohorts. Significant interaction tests were only obtained between each signature and the BLBC subtype for prediction of chemotherapy response or survival.

CONCLUSIONS

The proliferation signature predicts response and improved survival after chemotherapy, but only within BLBC. This highlights the clinical implications of TNBC heterogeneity, and suggests that future clinical trials focused on this phenotypic subtype should consider stratifying patients as having BLBC or not.

Surveillance of avian influenza A(H7N9) virus infection in humans and detection of the first imported human case in Taiwan, 3 April to 10 May 2013

On 3 April 2013, suspected and confirmed cases of influenza A(H7N9) virus infection became notifiable in the primary care sector in Taiwan, and detection of the virus became part of the surveillance of severe community-acquired pneumonia. On 24 April, the first imported case, reported through both surveillance systems, was confirmed in a man returning from China by sequencing from endotracheal aspirates after two negative throat swabs. Three of 139 contacts were ill and tested influenza A(H7N9)-negative.

Surveillance of avian influenza A(H7N9) virus infection in humans and detection of the first imported human case in Taiwan, 3 April to 10 May 2013

On 3 April 2013, suspected and confirmed cases of influenza A(H7N9) virus infection became notifiable in the primary care sector in Taiwan, and detection of the virus became part of the surveillance of severe community-acquired pneumonia. On 24 April, the first imported case, reported through both surveillance systems, was confirmed in a man returning from China by sequencing from endotracheal aspirates after two negative throat swabs. Three of 139 contacts were ill and tested influenza A(H7N9)-negative.

Abstract P6-05-02: Endocrine biomarkers in response to AR-inhibition with bicalutamide for the treatment of AR(+), ER/PR(−) metastatic breast cancer (MBC) (TBCRC011)

Background: Our group and others have identified a subset of ER/PR(−) breast cancers characterized by expression of the androgen receptor (AR) and androgen-dependent growth (Doane 2006). We conducted a proof-of-concept multicenter phase II study to test the efficacy of the AR-antagonist, bicalutamide for the treatment of AR(+) ER/PR(−) MBC (NCT00468715). Results of the primary endpoint, clinical benefit rate (CBR), were presented at ASCO (Gucalp 2012). Data for the impact of bicalutamide on circulating hormone levels in women are limited. Elevations in serum testosterone (T) and estradiol (E) have been observed for men treated with bicalutamide. We hypothesized comparable patterns of change in circulating endocrine markers in response to bicalutamide for women with MBC.

Methods: Patients (pts) with AR(+) (IHC ≥10%), ER/PR(−) (IHC <10%) MBC were eligible for treatment (tx) if ECOG performance status ≤2 and normal organ function regardless of menopausal status. There was no limit to prior tx except prior trastuzumab required if HER2(+). Tx consisted of bicalutamide 150mg orally daily in 28-day cycles (C). Toxicity assessed q4 weeks, response q12wks. Primary endpoint was CBR. Peripheral blood was collected for total and free T, E and sex hormone binding globulin (SHBG) at baseline, start of C2 (C2) and at end of study (EOS). Standard institutional assays were used. A Wilcoxon signed-rank test was done to compare baseline to C2 and EOS values.

Results: 26 patients with AR(+) ER/PR(−) MBC were treated on study. Evaluable number (n) of pts at baseline, C2 and EOS are 26, 26 and 19 respectively. Two pts remain on study. Menopausal status: pre=2, post=24. Baseline median total and free T and estradiol were consistent with expected norms, however a wide range was observed (Table). There were no significant differences observed for median free T, total T, E or SHBG between baseline and C2 or baseline and EOS. Changes in hormone levels could not be stratified by menopausal status or response to bicalutamide given small sample size. Given the wide range of baseline values, we examined the percent change for each endocrine biomarker from baseline to C2 and EOS. As shown in the Table, there was no difference in median percent change observed across time points for each biomarker.

Role of a putative tyrosine-O-sulfate receptor in the targeting and/or intracellular transport of tyrosine-sulfated proteins

By employing the affinity gel fraction technique, we have detected a 175 kDa tyrosine-O-sulfate (TyrS)-binding protein in sodium choleate extracts of the microsomal membrane fractions of bovine liver and pancreas, as well as canine liver and pancreas. Western blot analysis revealed the presence of the bovine liver TyrS-binding protein in complexes with tyrosine-sulfated proteins both in vivo and in vitro, suggesting the putative role of the former being the receptor for the latter. Using filter-grown Madin-Darby canine kidney (MDCK) cells as a model, it was demonstrated that the tyrosine-sulfated proteins synthesized were predominantly secreted into the apical medium. The results further indicate the production and differential polarized secretion of different sulfated forms of the two major secretory proteins produced by MDCK cells, fibronectin (FN) and an 80 kDa glycoprotein (gp 80), with their tyrosine-sulfated forms being predominantly secreted from the apical surface. Treatment of filter-grown MDCK cells with glycosylation inhibitors, swainsonine and 1-deoxymannojirimycin, appeared to enhance the apical secretion of tyrosine-sulfated FN and gp 80. A similar 175 kDa membrane-bound 'TyrS receptor', cross-reactive toward antiserum against the canine liver TyrS receptor, was shown to be present in MDCK cells. Pulse-chase experiments revealed its presence in complexes with newly synthesized FN and gp 80. A hypothetical model for TyrS residues serving as an apical targeting signal during the biosynthetic transport of tyrosine-sulfated proteins, as mediated by the TyrS receptor, in MDCK cells is proposed.

P5-11-03: Real-Time Imaging of Human Breast Tissue with Reflectance Confocal Microscopy: Correlation with Routine Pathology

Background: Near-infrared reflectance confocal microscopy (RCM) allows for immediate noninvasive 3-D optical sectioning of opaque objects, such as human tissue, without using the potentially destructive staining and fixing methods used with routine pathology. Recently, RCM has been used to differentiate between malignant and non-malignant dermatologic conditions. We hypothesize that this technique can be used to efficiently and reliably evaluate human breast tissue for the presence of malignancy without compromising the ability to perform routine immunohistochemical (IHC) analyses that accompany the diagnosis of invasive breast cancer.

Methods: 45 core needle breast biopsies (12mmx2mm) were collected under sonographic guidance. Biopsy specimens were immediately placed in phosphate buffered saline, injected with 5% acetic acid to enhance reflectivity of the nuclei, and imaged within 5–10 minutes. Digital images of the nuclear and cellular morphology from each intact specimen were acquired and catalogued within 1 hour of biopsy using the VivaCell 5000. Tissue samples were then formalin-fixed and sectioned for routine H&E evaluation or IHC assays. A board certified pathologist trained on 16 paired RCM images and H&E slides was given the blinded test set created from these 45 breast biopsies and asked to evaluate the RCM images for the presence of carcinoma. Preliminary evaluation was also done on 5 biopsy samples with known estrogen receptor (ER) and progesterone receptor (PR) status to determine the feasibility of assessing ER/PR on tissue treated with acetic acid for RCM.

Results: Routine H&E staining identified invasive carcinoma 10/45 biopsy samples (9 invasive ductal [IDC], 1 invasive lobular [ILC]); 2 specimens with IDC also contained ductal carcinoma in situ (DCIS). Evaluation of the RCM images led to the same diagnosis of invasive carcinoma vs. not in all 45 samples and correctly identified 6/9 IDC and 1/1 ILC. RCM correctly identified 1/2 specimens known to contain DCIS. RCM misclassified 3 IDC as DCIS (1), ILC (1), or lobular carcinoma in situ (1), and 1 DCIS as ILC. With respect to the determination of ER/PR status, RCM accurately assessed positivity for both receptors in all 5 specimens.

Conclusions: RCM is comparable to standard microscopy for the reliable identification of carcinoma, and the ability to evaluate breast tissue for malignancy using this technology allows for real-time pathology and may negate the need for repeat diagnostic biopsies to ensure adequate sampling for diagnosis. Importantly, the ability to perform routine IHC for ER/PR status after tissue processing for RCM appears to be preserved. Further evaluation applying the existing technology to a larger sample size for histopathologic correlation, assessment of ER/PR status, and evaluation for HER2 status are planned. These encouraging findings support interest in tailoring RCM for breast tissue to improve the ability to distinguish between invasive vs. in situ disease, and ductal vs. lobular histology.

Citation Information: Cancer Res 2011;71(24 Suppl):Abstract nr P5-11-03.

Association of vitamin D and antimicrobial peptide production during late-phase allergic responses in the lung

BACKGROUND

Vitamin D may play important roles in regulating immune responses and in defence against infectious diseases by effects on both innate and adaptive immune responses. Little is known regarding activation of vitamin D within airway tissues and its relationship to inflammation and antimicrobial responses.

OBJECTIVE

The objective of this study was to investigate the activation of vitamin D within the airways and to define relationships between vitamin D metabolites and measures of inflammatory and antimicrobial responses assessed by bronchoalveolar lavage (BAL) during late-phase responses following allergen challenge of allergic subjects.

METHODS

Segmental allergen challenge was performed with saline and allergen in 16 adult allergic subjects. BAL was performed in both saline and allergen-challenged sites 20-24 h. after challenge. Following extraction from BAL fluids, levels of 25-hydroxy-vitamin D (25(OH)D) and 1,25-dihydroxy-vitamin D (1,25(OH)(2)D) were assayed by specific radioimmunoassays. The cleavage product of cathelicidin, LL-37, was assayed by ELISA. Cellular constituents and albumin were measured.

RESULTS

Levels of vitamin D metabolites were increased in concentrated BAL fluids after allergen compared to saline challenge. Levels of 1,25(OH)(2)D increased from largely undetectable to 2.5 pm (median; range: 1-29.5; P = 0.005) while 25(OH)D increased from 3.2 (0.8-6.2) to 6.2 (1.5-184.9) nm (P = 0.0006). Levels of LL-37 increased from 2.1 (1.4-4.1) to 14.5 (2.2-106.7) ng/mL BAL (P = 0.0005). Levels of LL-37, 1,25(OH)(2)D, and 25(OH)D following allergen challenge were correlated with each other (P < 0.0001), cellular changes, and levels of albumin (P < 0.001).

CONCLUSIONS AND CLINICAL RELEVANCE

Levels of vitamin D metabolites, particularly 1,25(OH)(2)D, were low within the airways and increased after allergen challenge. The increases correlated with the magnitude of inflammation and increases in cathelicidin. Normalization to albumin suggested plasma exudation as a mechanism for the increases. The findings support a role for vitamin D in allergic and innate immune responses in the lung.

An experiment on the attrition of acid demineralized dentine in vitro

BACKGROUND

A laboratory investigation was designed to test the hypothesis that acids increase the rate of wear caused by attrition on dentine.

METHODS

Dentine sections from 10 teeth were polished, cleaned in an ultrasonic bath and divided into 8 equally sized areas. The occlusal tip of a tooth, placed vertically in a wear machine and loaded at 150 N, was moved against each dentine section for 5000 return strokes with artificial saliva acting as a lubricant. Each dentine section was divided into 8 sections and half randomly immersed in a 1% citric acid solution (pH 2.3) for 20 minutes. The wear regime produced 8 wear scars in total per dentine sample. The volume of each wear scar was measured using a contacting digitizing profilometer.

RESULTS

A total of 80 wear scars were produced with 40 treated with acid and 40 acting as controls. The mean for wear volume of the dentine scars with acid was 4.84 μm(3) (1.38) and for the non-acid surface 2.95 μm(3) (0.86). This difference was statistically significant (p < 0.05).

CONCLUSIONS

These results support the hypothesis that acids increase the rate of wear caused by attrition on dentine.

Pulmonary allergic responses augment interleukin-13 secretion by circulating basophils yet suppress interferon-alpha from plasmacytoid dendritic cells

BACKGROUND

Allergic inflammatory processes may have the capacity to propagate systemically through the actions of circulating leucocytes. Consequently, basophils from allergic individuals are often 'primed', as evidenced by their hyperresponsiveness in vitro. IFN-alpha secreted predominantly by plasmacytoid dendritic cells (pDCs), suppresses basophil priming for IL-13 production in vitro.

OBJECTIVE

This study sought in vivo correlates arising during experimental allergen challenge that support an 'axis-interplay' between basophils and pDCs.

METHODS

Using segmental allergen challenge (SAC) in the lung, the immune responses of both cell types from the blood were investigated in volunteers (n=10) before and 24 h after allergen exposure. These responses were then correlated with inflammatory parameters measured in bronchoalveolar lavage fluids (BALF).

RESULTS

In the blood, SAC significantly augmented IL-13 secretion by basophils induced by IL-3 (P=0.009), yet reduced IFN-alpha secreted by pDCs stimulated with CpG (P=0.018). Both parameters were negatively correlated (P=0.0015), at least among those subjects that secreted the latter. Circulating basophil IL-13 responses further correlated with post-SAC bronchoalveolar lavage (BAL) parameters including IL-13 protein (P=0.04), basophil (P=0.051), eosinophil (P=0.0018), and total cell counts (P<0.003). Basophil and IL-13 levels in BAL correlated likewise (P=0.0002).

CONCLUSIONS

These results support a mechanism of immune regulation whereby an allergen reduces innate immune responses and IFN-alpha production by pDCs, resulting in an enhanced inflammation and basophil cytokine production at sites of allergen exposure.

Time-dependent gene expression changes in a xenograft model of docetaxel cytotoxicity

Abstract #2009

Background: The antitumor effects of docetaxel primarily result from mitotic arrest with impaired cell proliferation. However, the key molecular pathways affected by this agent are largely unknown, and alternative mechanisms of action are likely. Knowledge of these pathways and mechanisms will guide the rational design of more effective treatment schedules and drug combinations. We hypothesize that the molecular analysis of treated human breast cancer xenografts will allow for an evaluation of time-dependent alterations in the expression of genes involved in key cell processes related to docetaxel cytotoxicity.&#x2028; Methods: Gene expression profiles were generated from MDA-MB-231 breast cancer xenografts after a single intravenous dose of docetaxel. Tumors were harvested at 1, 6, 12, and 24 hours (hr) after drug administration; a control group of untreated tumors (0hr) was included. RNA was extracted, assessed for concentration and quality, and used to generate probes for hybridization to the Affymetrix Human U133A oligonucleotide gene chip. Gene expression levels were explored with MAS 5.0 software and in-house algorithms. Subsequent data analysis included cross-experiment normalization, differential gene selection, gene clustering, and peak-time latency analysis. Consistency between the in-house in vivo dataset and a public in vitro dataset (Oncogene. 26:2902) was investigated.&#x2028; Results: 5087 responsive genes were selected using both SAM and EDGE. Among these genes, 3009 were down-regulated and 190 were up-regulated by at least 2-fold. Compared with 0h, the most significant changes in gene expression took place 6h after drug administration (2023 down-regulated and 655 up-regulated genes). Gene clustering revealed the sequence of up-regulation or down-regulation of different gene groups. Apoptosis-related genes and cell cycle-related genes in various stages showed different time latency patterns. For example, most of the M phase genes (total of 65 from Gene Ontology) were down-regulated 6h and 12h after docetaxel treatment and recovered by 24h. In addition, 1489 of 3009 down-regulated genes and 55 of 190 up-regulated genes were found in a public in vitro dataset; these datasets share 282 down-regulated genes (18.9%) and 17 up-regulated genes (30.9%).&#x2028; Discussion: Analysis of serial gene expression changes in docetaxel sensitive breast cancer xenografts demonstrates that time-dependent changes in specific cell processes occur. These findings will be expanded to distinguish between early and late molecular mechanisms of docetaxel responsiveness. A parallel xenograft model of docetaxel nonresponsiveness is also under investigation. In this way, time-dependent pathways will be defined that will include predictive biomarkers and lead to the development of novel agents to overcome resistance or enhance sensitivity to docetaxel, one of the most effective agents available for the treatment of breast cancer.

Citation Information: Cancer Res 2009;69(2 Suppl):Abstract nr 2009.

Effects of inhaled human insulin on airway lining fluid composition in adults with diabetes

Inhaled human insulin (Exubera (human insulin of rDNA origin) Inhalation Powder) causes small, early and reversible changes in pulmonary function in subjects with diabetes mellitus. The present study assessed whether changes occur in cellular and soluble constituents of airway lining fluid consistent with inflammation as a possible cause for Exubera-associated lung function alterations. Two 31-week, open-label, sequential design phase 2 studies were conducted, one with 20 subjects with type 1 and one with 24 subjects with type 2 diabetes. After run-in, all subjects received subcutaneous insulin for 12 weeks, followed after 1 week by 12 weeks of Exubera. Bronchoalveolar lavage fluid cell counts and protein constituents were determined at baseline, after 12 weeks of subcutaneous insulin and after 12 weeks of Exubera. Baseline cellular and soluble constituents of lavage fluid were similar to those reported for nondiabetic adults. Exubera produced no consistent clinically or statistically significant changes in total or differential lavage fluid cell counts or protein concentrations, even though Exubera-associated changes in pulmonary function are known to be fully manifest within 12 weeks. Therefore, 12 weeks of Exubera treatment is not associated with evidence of pulmonary inflammation. The treatment effects on lung function observed in Exubera trials are not caused by lung inflammation.

The characteristics of neuronal injury in a static compression model of spinal cord injury in adult rats

Studies of spinal cord injury using contusion (impact) injury paradigms have shown that neuronal death is an acute event that is largely over within 24 h. However, much less is known about cell death following compression injury, despite compression being a key component of natural spinal injuries. We have therefore used neuronal nuclei (NeuN) immunostaining to examine the spatiotemporal pattern of neuronal loss after static compression injury in adult rats. 3D reconstruction was used to reveal the full effect of the injury. Neuronal loss at the injury epicentre, assessed by NeuN immunostaining, amounted to 44% at 1 day but increased to 73% at 3 days and 81% at 1 month. Neuronal loss was also seen 5 mm rostral and caudal to the epicentre, but was not significant until 3 days. NeuN loss was greatest in the ventral horns and in the intermediate grey matter, with the lateral dorsal horns relatively spared. Cystic cavities formed after injury, but were not evident until 4 weeks and were small in size. In contrast to the slow profile of neuronal loss, the compression injury also evoked a transient expression of activating transcription factor-3 (ATF3) and activated c-Jun in neurons. ATF3 expression peaked at 3 days and declined at 7 days. Our spatiotemporal analysis of compression injury shows that neuronal loss is much more protracted than in contusion injury, and highlights the potential for neuroprotective strategies. This study is also the first indication of ATF3 involvement in spinal cord injury.

Spinal cord compression and dorsal root injury cause up-regulation of activating transcription factor-3 in large-diameter dorsal root ganglion neurons

Spinal cord injury causes damage to ascending and descending tracts, as well as to local circuits, but relatively little is known about the effect of such injury on sensory neurons located within adjoining ganglia. We have therefore used immunocytochemistry for activating transcription factor-3 (ATF3), a sensitive marker of axonal damage, in order to examine the effects of spinal cord injury in rats on dorsal root ganglion (DRG) neurons. A 50-g static compression injury applied to the dorsal surface of the T12 thoracic spinal cord led to an up-regulation of ATF3 that was maximal at 1 day and affected 12-14% of DRG neurons in ganglia caudal to the injury (T13-L3). A similar response was seen after a T12 hemisection that transected the dorsal columns except that compression injury, but not hemisection, also evoked ATF3 expression in ganglia just rostral to the injury (T10, T11). ATF3 was up-regulated exclusively in DRG neurons that were of large diameter and immunoreactive for heavy neurofilament. Small-diameter cells, including the population that binds the lectin Grifffonia simplicifolia IB4, did not express ATF3 immunoreactivity. A similar pattern of ATF3 expression was induced by dorsal rhizotomy. The data show for the first time that ATF3 is up-regulated after spinal cord and dorsal root injury, but that this up-regulation is confined to the large-diameter cell population.

Quantitative structure-toxicity relationships (QSTRs): a comparative study of various non linear methods. General regression neural network, radial basis function neural network and support vector machine in predicting toxicity of nitro- and cyano- aromatics to Tetrahymena pyriformis

Prediction of toxicity of 203 nitro- and cyano-aromatic chemicals to Tetrahymena pyriformis was carried out by radial basis function neural network, general regression neural network and support vector machine, in non-linear response surface methodology. Toxicity was predicted from hydrophobicity parameter (log Kow) and maximum superdelocalizability (Amax). Special attention was drawn to prediction ability and robustness of the models, investigated both in a leave-one-out and 10-fold cross validation (CV) processes. The influence that the corresponding changes in the learning sets during these CV processes could have on a common external test set including 41 compounds was also examined. This allowed us to establish the stability of the models. The non linear results slightly outperform (as expected) multilinear relationships (MLR) and also favourably compete with various other non linear approaches recently proposed by Ren (J. Chem. Inf. Comput. Sci., 43 1679 (2003)).

Support vector machines-based quantitative structure-property relationship for the prediction of heat capacity

The support vector machine (SVM), as a novel type of learning machine, for the first time, was used to develop a Quantitative Structure-Property Relationship (QSPR) model of the heat capacity of a diverse set of 182 compounds based on the molecular descriptors calculated from the structure alone. Multiple linear regression (MLR) and radial basis function networks (RBFNNs) were also utilized to construct quantitative linear and nonlinear models to compare with the results obtained by SVM. The root-mean-square (rms) errors in heat capacity predictions for the whole data set given by MLR, RBFNNs, and SVM were 4.648, 4.337, and 2.931 heat capacity units, respectively. The prediction results are in good agreement with the experimental value of heat capacity; also, the results reveal the superiority of the SVM over MLR and RBFNNs models.

Comparative study of QSAR/QSPR correlations using support vector machines, radial basis function neural networks, and multiple linear regression

Support vector machines (SVMs) were used to develop QSAR models that correlate molecular structures to their toxicity and bioactivities. The performance and predictive ability of SVM are investigated and compared with other methods such as multiple linear regression and radial basis function neural network methods. In the present study, two different data sets were evaluated. The first one involves an application of SVM to the development of a QSAR model for the prediction of toxicities of 153 phenols, and the second investigation deals with the QSAR model between the structures and the activities of a set of 85 cyclooxygenase 2 (COX-2) inhibitors. For each application, the molecular structures were described using either the physicochemical parameters or molecular descriptors. In both studied cases, the predictive ability of the SVM model is comparable or superior to those obtained by MLR and RBFNN. The results indicate that SVM can be used as an alternative powerful modeling tool for QSAR studies.

Prediction of the tissue/blood partition coefficients of organic compounds based on the molecular structure using least-squares support vector machines

The accurate nonlinear model for predicting the tissue/blood partition coefficients (PC) of organic compounds in different tissues was firstly developed based on least-squares support vector machines (LS-SVM), as a novel machine learning technique, by using the compounds' molecular descriptors calculated from the structure alone and the composition features of tissues. The heuristic method (HM) was used to select the appropriate molecular descriptors and build the linear model. The prediction result of the LS-SVM model is much better than that obtained by HM method and the prediction values of tissue/blood partition coefficients based on the LS-SVM model are in good agreement with the experimental values, which proved that nonlinear model can simulate the relationship between the structural descriptors, the tissue composition and the tissue/blood partition coefficients more accurately as well as LS-SVM was a powerful and promising tool in the prediction of the tissue/blood partition behaviour of compounds. Furthermore, this paper provided a new and effective method for predicting the tissue/blood partition behaviour of the compounds in the different tissues from their structures and gave some insight into structural features related to the partition process of the organic compounds in different tissues.

Diagnosing anorexia based on partial least squares, back propagation neural network, and support vector machines

Support vector machine (SVM), as a novel type of learning machine, for the first time, was used to develop a predictive model for early diagnosis of anorexia. It was based on the concentration of six elements (Zn, Fe, Mg, Cu, Ca, and Mn) and the age extracted from 90 cases. Compared with the results obtained from two other classifiers, partial least squares (PLS) and back-propagation neural network (BPNN), the SVM method exhibited the best whole performance. The accuracies for the test set by PLS, BPNN, and SVM methods were 52%, 65%, and 87%, respectively. Moreover, the models we proposed could also provide some insight into what factors were related to anorexia.

Application of support vector machine (SVM) for prediction toxic activity of different data sets

As a new method, support vector machine (SVM) were applied for prediction of toxicity of different data sets compared with other two common methods, multiple linear regression (MLR) and RBFNN. Quantitative structure-activity relationships (QSAR) models based on calculated molecular descriptors have been clearly established. Among them, SVM model gave the highest q(2) and correlation coefficient R. It indicates that the SVM performed better generalization ability than the MLR and RBFNN methods, especially in the test set and the whole data set. This eventually leads to better generalization than neural networks, which implement the empirical risk minimization principle and may not converge to global solutions. We would expect SVM method as a powerful tool for the prediction of molecular properties.

QSAR study of natural, synthetic and environmental endocrine disrupting compounds for binding to the androgen receptor

A large data set of 146 natural, synthetic and environmental chemicals belonging to a broad range of structural classes have been tested for their relative binding affinity (expressed as log (RBA)) to the androgen receptor (AR). These chemicals commonly termed endocrine disrupting compounds (EDCs) present a variety of adverse effects in humans and animals. As assays for binding affinity remains a time-consuming task, it is important to develop predictive methods. In this work, quantitative structure-activity relationships (QSARs) were determined using three methods, multiple linear regression (MLR), radical basis function neural network (RBFNN) and support vector machine (SVM). Five descriptors, accounting for hydrogen-bonding interaction, distribution of atomic charges and molecular branching degree, were selected from a heuristic method to build predictive QSAR models. Comparison of the results obtained from three models showed that the SVM method exhibited the best overall performances, with a RMS error of 0.54 log (RBA) units for the training set, 0.59 for the test set, and 0.55 for the whole set. Moreover, six linear QSAR models were constructed for some specific families based on their chemical structures. These predictive toxicology models, should be useful to rapidly identify potential androgenic endocrine disrupting compounds.

Study of probabilistic neural networks to classify the active compounds in medicinal plants

Probabilistic neural networks (PNNs) were utilized for the classifications of 102 active compounds from diverse medicinal plants with anticancer activity against human rhinopharyngocele cell line KB. Molecular descriptors calculated from structure alone were used to represent molecular structures. A subset of the calculated descriptors selected using factor correlation analysis and forward stepwise regression was used to construct the prediction models. Linear discriminant analysis (LDA) was also utilized to construct the classification model to compare the results with those obtained by PNNs. The accuracy of the training set, the cross-validation set, and the test set given by PNNs and LDA were 100, 92.3, 90.9% and 71.8, 92.3, 54.5%, respectively, which indicated that the results obtained by PNNs agree well with the experimental values of these compounds and also revealed the superiority of PNNs over LDA approach for the classification of anticancer activities of compounds. The models built in this work would be of potential help in the design of novel and more potent anticancer agents.

QSAR and classification models of a novel series of COX-2 selective inhibitors: 1,5-diarylimidazoles based on support vector machines

The support vector machine, which is a novel algorithm from the machine learning community, was used to develop quantitation and classification models which can be used as a potential screening mechanism for a novel series of COX-2 selective inhibitors. Each compound was represented by calculated structural descriptors that encode constitutional, topological, geometrical, electrostatic, and quantum-chemical features. The heuristic method was then used to search the descriptor space and select the descriptors responsible for activity. Quantitative modelling results in a nonlinear, seven-descriptor model based on SVMs with root mean-square errors of 0.107 and 0.136 for training and prediction sets, respectively. The best classification results are found using SVMs: the accuracy for training and test sets is 91.2% and 88.2%, respectively. This paper proposes a new and effective method for drug design and screening.

3D QSAR studies on antimalarial alkoxylated and hydroxylated chalcones by CoMFA and CoMSIA

The 3D QSAR analyses of antimalarial alkoxylated and hydroxylated chalcones were first conducted by Comparative molecular field analysis (CoMFA) and Comparative similarity indices analysis (CoMSIA) to determine the factors required for the activity of these compounds. Satisfactory results were obtained after performing a leave-one-out (LOO) cross-validation study with cross-validation q(2) and conventional r(2) values of 0.740 and 0.972 by the CoMFA model, 0.714 and 0.976 by the CoMSIA model, respectively. The results provided the tools for predicting the affinity of related compounds, and for guiding the design and synthesis of novel and more potent antimalarial agents.