Over the past few years, establishment and adaptation of cell-based assays for drug development and testing has become an important topic in high-throughput screening (HTS). Most new assays are designed to rapidly detect specific cellular effects reflecting action at various targets. However, although more complex than cell-free biochemical test systems, HTS assays using monolayer or suspension cultures still reflect a highly artificial cellular environment and may thus have limited predictive value for the clinical efficacy of a compound. Today's strategies for drug discovery and development, be they hypothesis free or mechanism based, require facile, HTS-amenable test systems that mimic the human tissue environment with increasing accuracy in order to optimize preclinical and preanimal selection of the most active molecules from a large pool of potential effectors, for example, against solid tumors. Indeed, it is recognized that 3-dimensional cell culture systems better reflect the in vivo behavior of most cell types. However, these 3-D test systems have not yet been incorporated into mainstream drug development operations. This article addresses the relevance and potential of 3-D in vitro systems for drug development, with a focus on screening for novel antitumor drugs. Examples of 3-D cell models used in cancer research are given, and the advantages and limitations of these systems of intermediate complexity are discussed in comparison with both 2-D culture and in vivo models. The most commonly used 3-D cell culture systems, multicellular spheroids, are emphasized due to their advantages and potential for rapid development as HTS systems. Thus, multicellular tumor spheroids are an ideal basis for the next step in creating HTS assays, which are predictive of in vivo antitumor efficacy. (Journal of Biomolecular Screening 2004:273-285)

To detect novel Wnt-pathway genes involved in tumourigenesis, this study analysed the RNA expression levels of 40 genes of the Wnt pathway by chip hybridization of microdissected matched pairs of 54 primary prostate carcinomas. Eleven genes showed greater than two-fold differential expression in at least 10% of prostate cancers. Three of these genes encode extracellular components of the Wnt pathway (WNT2, WIF1, SFRP4); two are receptors (FZD4, FZD6); two belong to the intracellular signal cascade (DVL1, PPP2CB); one regulates transcription (TCF4); and three represent genes regulated by this pathway (CCND2, CD44, MYC). While SFRP4, FZD4, FZD6, DVL1, TCF4, and MYC are up-regulated, WIF1, WNT2, PPP2CB, CCND2, and CD44 are down-regulated in certain prostate cancer patients. Wnt inhibitory factor 1 (WIF1) and secreted frizzled related protein (SFRP4) showed the most significant aberrant expression at the RNA level. WIF1 was down-regulated in 64% of primary prostate cancers, while SFRP4 was up-regulated in 81% of the patients. Immunohistochemical analysis using a polyclonal antibody revealed strong cytoplasmic perinuclear WIF1 expression in normal epithelial cells of the prostate, breast, lung, and urinary bladder. Strong reduction of WIF1 protein expression was found in 23% of prostate carcinomas, but also in 60% of breast, 75% of non-small cell lung (NSCLC), and 26% of bladder cancers analysed. No significant association between WIF1 down-regulation and tumour stage or grade was observed for prostate, breast or non-small cell lung carcinomas, indicating that loss of WIF1 expression may be an early event in tumourigenesis in these tissues. However, down-regulation of WIF1 correlated with higher tumour stage in urinary bladder tumours (pTa versus pT1-pT4; p = 0.038).

Multifocal hyperplasia of interstitial cells of Cajal (ICC hyperplasia) is a precursor of hereditary gastrointestinal stromal tumors (GISTs) in patients with germline mutations of c-KIT or PDGFRA, but precursor lesions of sporadic GISTs have not been defined yet. Small hyalinizing stromal tumors of the proximal stomach (referred to in this study as GIST tumorlets) were collected prospectively from 98 consecutive autopsies and additional cases were retrieved from surgical pathology files (total n=57). GIST tumorlets were grossly detectable in 22.5% consecutive autopsies performed in individuals older than 50 years. All lesions were located in the cardia, fundus, or proximal body, and ranged in size from 1 to 10 mm (4 mm). Similar lesions were not detected in the antrum, duodenum, and the remainder of the bowel. Histologically, the spindle cell subtype comprised all cases, with hyalinization and calcification in 57% of cases. The spindle cells were immunohistochemically positive for vimentin, CD117, and CD34. Twenty-four cases yielded sufficient DNA for subsequent molecular analysis, which showed c-KIT mutations in 11 cases (46%) and PDGFRA mutations in 1 case (4%). Sporadic GIST tumorlets of the proximal stomach are common in the general population over the age of 50 years and frequently show somatic c-KIT mutations. GIST tumorlets probably represent the grossly recognizable counterpart of sporadic ICC hyperplasia caused by somatic c-KIT or PDGFRA mutations. Early hyalinization and calcification seems to confer limited growth potential, and complete regression of such lesions is common. GIST tumorlets likely represent preclinical (preneoplastic) lesions that need additional stimuli to evolve into clinical GISTs, raising the possibility of a hyperplasia-neoplasia sequence in the development of sporadic GISTs.

Malignant cells identified regularly in the blood shed during tumor surgery and different from circulating tumor cells are of concern, since at the surgical site they may cause local tumor recurrence, or in the salvaged blood they may cause hematogenic metastasis after retransfusion. Therefore, the contraindication of intraoperative autotransfusion in tumor surgery is strongly supported, and a review of surgical procedures and adjuvant therapy may be indicated, as the passage of the identified cells to the shed blood is yet unknown.

No abstract