Janowska-Wieczorek et al., pp.752–760 Fragments of the plasma membrane shed by cells are suspected to participate in intercellular signaling. These fragments, called microvesicles, contain proteins similar, but not identical, to those in the membrane of the originating cells. Platelets release microvesicles when stimulated by agonists or exposed to high-stress shear forces, and these platelet-derived microvesicles (PMV) regulate proliferation, survival, and adhesion of human normal and malignant hematopoietic cells. Experimental metastases can be reduced by decreasing the blood platelet concentration, and increased by transfusion of platelets. Janowska-Wieczorek and colleagues suspected that PMV could spur lung cancer cells to grow and metastasize. They investigated the transfer of platelet-derived surface molecules to lung cancer cells in vitro as well as the direct biologic effect of PMV on the cancer cells. They report that PMV interact with lung cancer cells and increase their metastatic and angiogenic potential. Not only do the microvesicles chemoattract the cancer cells, they stimulate various signaling pathways, including some involved in proliferation and vascularization. In addition, the authors showed that in mice, lung carcinoma cells covered with PMV spawned more metastatic loci in the lungs and spread to the bone marrow in greater numbers than did cancer cells without PMV. These results help to explain the aggravating effect of platelets on cancer. The authors suggest that PMV play a key role in mediating communication between platelets and cancer cells, but that more work will be critical to determine exactly which components do the talking, and potentially exploit them for new therapeutic approaches. Chin et al., pp.789–797 Although head and neck cancer can be successfully treated when detected early, the survival rate is not high, because most cases are only discovered after the disease has advanced. Treatment usually incorporates surgery, radiotherapy, and chemotherapy. Although surgery has the best outcome, its unpleasant side effects make it undesirable if the disease is likely to recur despite treatment. Having a marker that identified which cancers were most likely to recur would greatly help physicians choose the most appropriate course of treatment. To find such a marker, Chin and colleagues compared the expression of more than 13,000 genes in cancerous and normal tissue. They found that, among others, the protein SPARC (osteonectin) is consistently elevated in head and neck cancers. SPARC is known to be overexpressed in other tumors as well, but has never been associated with head and neck cancer previously. The authors also examined gene expression in relation to clinical outcome in a number of patients. SPARC turned out to be a strong indicator of a short disease-free interval and poor prognosis. The association became even stronger when combined with two other extracellular matrix proteins identified, PAI-1 and uPA. This study marks that matched tissue samples have been analyzed for prognostic markers, and identified markers have been verified with clinical data and prognosis. These markers could potentially help determine whether a conservative course of treatment is warranted, and may even serve as early warning indicators of recurrence. Ding et al., pp.803–810 Nonsteroidal anti-inflammatory drugs (NSAIDs) have shown some anticancer activity in many types of human cancers. These drugs act on cyclooxygenases (COX), and because COX-2 is often overexpressed in cancer cells, NSAIDs are thought to impair cancer growth by targeting COX-2. Ding and colleagues investigated the role of one such drug, celecoxib, in pre-malignant and malignant oral cell culture. Previous evidence has suggested that celecoxib could prompt apoptosis through pathways independent of COX-2, so the researchers evaluated second-generation celecoxib derivatives that lacked COX-2 inhibitory activity. They report that cells exposed to these derivatives pause in the G2/M phase of the cell cycle. The derivatives also induced apoptosis with greater potency than celecoxib did. Although earlier work with prostate cancer cells had described a pathway by which these derivatives boosted apoptosis, the current work shows that, in oral cells, derivative-induced apoptosis occurs independently of that pathway. The results suggest that celecoxib derivatives can take advantage of a variety of molecular mechanisms to inhibit growth and promote apoptosis in many cell types. Although results obtained in vitro may not always reflect tumor responses in vivo, the current work provides a reasonable model for designing celecoxib compounds with potential for use in chemoprevention and chemotherapy.