Preclinical models for neuroblastoma: establishing a baseline for treatment

Background

Preclinical models of pediatric cancers are essential for testing new chemotherapeutic combinations for clinical trials. The most widely used genetic model for preclinical testing of neuroblastoma is the TH-MYCN mouse. This neuroblastoma-prone mouse recapitulates many of the features of human neuroblastoma. Limitations of this model include the low frequency of bone marrow metastasis, the lack of information on whether the gene expression patterns in this system parallels human neuroblastomas, the relatively slow rate of tumor formation and variability in tumor penetrance on different genetic backgrounds. As an alternative, preclinical studies are frequently performed using human cell lines xenografted into immunocompromised mice, either as flank implant or orthtotopically. Drawbacks of this system include the use of cell lines that have been in culture for years, the inappropriate microenvironment of the flank or difficult, time consuming surgery for orthotopic transplants and the absence of an intact immune system.

Principal Findings

Here we characterize and optimize both systems to increase their utility for preclinical studies. We show that TH-MYCN mice develop tumors in the paraspinal ganglia, but not in the adrenal, with cellular and gene expression patterns similar to human NB. In addition, we present a new ultrasound guided, minimally invasive orthotopic xenograft method. This injection technique is rapid, provides accurate targeting of the injected cells and leads to efficient engraftment. We also demonstrate that tumors can be detected, monitored and quantified prior to visualization using ultrasound, MRI and bioluminescence. Finally we develop and test a “standard of care” chemotherapy regimen. This protocol, which is based on current treatments for neuroblastoma, provides a baseline for comparison of new therapeutic agents.

Significance

The studies suggest that use of both the TH-NMYC model of neuroblastoma and the orthotopic xenograft model provide the optimal combination for testing new chemotherapies for this devastating childhood cancer.     

Structure of the carboxypeptidase Y inhibitor IC in complex with the cognate proteinase reveals a novel mode of the proteinase-protein inhibitor interaction

Carboxypeptidase Y (CPY) inhibitor, IC, shows no homology to any other known proteinase inhibitors and rather belongs to the phosphatidylethanolamine-binding protein (PEBP) family. We report here on the crystal structure of the IC-CPY complex at 2.7 A resolution. The structure of IC in the complex with CPY consists of one major beta-type domain and a N-terminal helical segment. The structure of the complex contains two binding sites of IC toward CPY, the N-terminal inhibitory reactive site (the primary CPY-binding site) and the secondary CPY-binding site, which interact with the S1 substrate-binding site of CPY and the hydrophobic surface flanked by the active site of the enzyme, respectively. It was also revealed that IC had the ligand-binding site, which is conserved among PEBPs and the putative binding site of the polar head group of phospholipid. The complex structure and analyses of IC mutants for inhibitory activity and the binding to CPY demonstrate that the N-terminal inhibitory reactive site is essential both for inhibitory function and the complex formation with CPY and that the binding of IC to CPY constitutes a novel mode of the proteinase-protein inhibitor interaction. The unique binding mode of IC toward the cognate proteinase provides insights into the inhibitory mechanism of PEBPs toward serine proteinases and into the specific biological functions of IC belonging to the PEBP family as well.     

Enhanced reactivity of Rhizopus oryzae lipase displayed on yeast cell surfaces in organic solvents: potential as a whole-cell biocatalyst in organic solvents

Immobilization of enzymes on some solid supports has been used to stabilize enzymes in organic solvents. In this study, we evaluated applications of genetically immobilized Rhizopus oryzae lipase displayed on the cell surface of Saccharomyces cerevisiae in organic solvents and measured the catalytic activity of the displayed enzyme as a fusion protein with alpha-agglutinin. Compared to the activity of a commercial preparation of this lipase, the activity of the new preparation was 4.4 x 10(4)-fold higher in a hydrolysis reaction using p-nitrophenyl palmitate and 3.8 x 10(4)-fold higher in an esterification reaction with palmitic acid and n-pentanol (0.2% H2O). Increased enzyme activity may occur because the lipase displayed on the yeast cell surface is stabilized by the cell wall. We used a combination of error-prone PCR and cell surface display to increase lipase activity. Of 7,000 colonies in a library of mutated lipases, 13 formed a clear halo on plates containing 0.2% methyl palmitate. In organic solvents, the catalytic activity of 5/13 mutants was three- to sixfold higher than that of the original construct. Thus, yeast cells displaying the lipase can be used in organic solvents, and the lipase activity may be increased by a combination of protein engineering and display techniques. Thus, this immobilized lipase, which is more easily prepared and has higher activity than commercially available free and immobilized lipases, may be a practical alternative for the production of esters derived from fatty acids.     

Structure-activity relationships of 2alpha-substituted androstenedione analogs as aromatase inhibitors and their aromatization reactions

Aromatase catalyzes the conversion of androstenedione (1a, AD) to estrone through three sequential oxygenations of the 19-methyl group. To gain insight into the spatial nature of the AD binding (active) site of aromatase in relation to the catalytic function of the enzyme, we tested for the ability of 2alpha-substituted (halogeno, alkyl, hydroxy, and alkoxy) ADs (1b-1i) to inhibit aromatase in human placental microsomes as well as their ability to serve as a substrate for the enzyme. All of the steroids inhibited the enzyme in a competitive manner with the apparent K(i)'s ranging from 45 to 1150 nM. 2alpha-Halogeno (F, Cl, and Br) and 2alpha-alkyl (CH3 and CH2CH3) steroids 1b-1f were powerful to good inhibitors (Ki=45-171 nM) whereas steroids 1g-1i, having an oxygen function (hydroxy or alkoxy) at C-2alpha, were poor inhibitors (Ki=670-1150 nM). Aromatization of some of the steroids with placental microsomes was analyzed by gas chromatography-mass spectrometry, indicating that the aromatization rate of the bromide 1d was about two-fold that of the natural substrate AD and that of 2alpha-methoxide 1h was similar to that of AD. Kinetic analysis of the aromatization of androgens revealed that a good substrate was not essentially a good inhibitor for aromatase.     

Serinc, an activity-regulated protein family, incorporates serine into membrane lipid synthesis

Cell membranes contain various transporter proteins, some of which are responsible for transferring amino acids across membrane. In this study, we report another class of carrier proteins, termed Serinc1-5, that incorporates a polar amino acid serine into membranes and facilitates the synthesis of two serine-derived lipids, phosphatidylserine and sphingolipids. Serinc is a unique protein family that shows no amino acid homology to other proteins but is highly conserved among eukaryotes. The members contain 11 transmembrane domains, and rat Serinc1 protein co-localizes with lipid biosynthetic enzymes in endoplasmic reticulum membranes. A Serinc protein forms an intracellular complex with key enzymes involved in serine and sphingolipid biosyntheses, and both functions, serine synthesis and membrane incorporation, are linked to each other. In the rat brain, expression of Serinc1 and Serinc2 mRNA was rapidly up-regulated by kainate-induced seizures in neuronal cell layers of the hippocampus. In contrast, myelin throughout the brain is enriched with Serinc5, which was down-regulated in the hippocampus by seizures. These results indicate a novel mechanism linking neural activity and lipid biosynthesis.     

Smad3 mediates TGF-beta1 induction of VEGF production in lung fibroblasts

Transforming growth factor-beta1 (TGF-beta1) is a key factor in a variety of physiological and pathological processes. Vascular endothelial growth factor (VEGF) is a key angiogenic factor, and vascular change is one of the features of airway remodeling. We examined the effect of TGF-beta1 on VEGF production by fibroblasts from mice lacking expression of Smad2 or Smad3 as well as human lung fibroblasts treated with or without Smad2 or Smad3 siRNA. TGF-beta1 stimulated VEGF production by fibroblasts from Smad2 deficient animals and wildtype animals. In contrast, TGF-beta1 did not affect VEGF production by fibroblasts from Samd3 deficient mice. Similarly, TGF-beta1 failed to stimulate VEGF production by HFL-1 cells treated with Samd3 siRNA but significantly increased VEGF production by the cells treated with Smad2 siRNA. These result suggest that TGF-beta1 stimulation of VEGF production by fibroblasts is regulated by Smad3 but not by Smad2 signaling.     

Optimal site-specific PEGylation of mutant TNF-alpha improves its antitumor potency

Recently, we created a lysine-deficient mutant tumor necrosis factor-alpha [mTNF-alpha-Lys(-)] with full bioactivity in vitro compared with wild-type TNF-alpha (wTNF-alpha), and site-specific PEGylation of mTNF-alpha-Lys(-) was found to selectively enhance its in vivo antitumor activity. In this study, we attempted to optimize this PEGylation of mTNF-alpha-Lys(-) to further improve its therapeutic potency. mTNF-alpha-Lys(-) was site-specifically modified at its N-terminus with linear polyethylene glycol (LPEG) or branched PEG (BPEG). While randomly mono-PEGylated wTNF-alpha (ran-LPEG5K-wTNF-alpha) with 5 kDa of LPEG (LPEG5K) had about only 4% in vitro bioactivity of wTNF-alpha, mono-PEGylated mTNF-alpha-Lys(-) [sp-PEG-mTNF-alpha-Lys(-)] with LPEG5K, LPEG20K, BPEG10K, and BPEG40K had 82%, 58%, 93%, and 65% bioactivities of mTNF-alpha-Lys(-), respectively. sp-LPEG-mTNF-alpha-Lys(-) and sp-BPEG10K-mTNF-alpha-Lys(-) had much superior antitumor activity to those of both unmodified TNF-alphas and ran-LPEG5K-wTNF-alpha, though sp-BPEG40K-mTNF-alpha-Lys(-) did not show in vivo antitumor activity. Thus, the molecular shape and weight of PEG may strongly influence the in vivo antitumor activity of sp-PEG-mTNF-alpha-Lys(-).