Efficient nucleus detector in histopathology images.

In traditional cancer diagnosis, (histo)pathological images of biopsy samples are visually analysed by pathologists. However, this judgment is subjective and leads to variability among pathologists. Digital scanners may enable automated objective assessment, improved quality and reduced throughput time. Nucleus detection is seen as the corner stone for a range of applications in automated assessment of (histo)pathological images. In this paper, we propose an efficient nucleus detector designed with machine learning. We applied colour deconvolution to reconstruct each applied stain. Next, we constructed a large feature set and modified AdaBoost to create two detectors, focused on different characteristics in appearance of nuclei. The proposed modification of AdaBoost enables inclusion of the computational cost of each feature during selection, thus improving the computational efficiency of the resulting detectors. The outputs of the two detectors are merged by a globally optimal active contour algorithm to refine the border of the detected nuclei. With a detection rate of 95% (on average 58 incorrectly found objects per field-of-view) based on 51 field-of-view images of Her2 immunohistochemistry stained breast tissue and a complete analysis in 1 s per field-of-view, our nucleus detector shows good performance and could enable a range of applications in automated assessment of (histo)pathological images.

The use of hydrophobins to functionalize surfaces.

The physiochemical nature of surfaces can be changed by small proteins which are secreted by filamentous fungi. These proteins, called hydrophobins, are characterized by the presence of eight conserved cysteine residues and a typical hydropathy pattern. Upon contact with a hydrophilic-hydrophobic interface they self-assemble into highly insoluble amphipathic membranes. As a result, hydrophobic surfaces become hydrophilic and vice versa. Genetic engineering of hydrophobins was used to study structure-function relationships. In addition, engineered hydrophobins were constructed to increase the biocompatibility of surfaces. The glycosylated N-terminal region of the mature SC3 hydrophobin was deleted and the cell-binding domain of human fibronectin was introduced at the N-terminus. The gross properties of the hydrophobins were not affected. However, the physiochemical properties of the hydrophilic side of the assembled protein did change. Growth of fibroblasts on Teflon could be improved by coating the solid with the engineered hydrophobins. Thus, by changing the N-terminal part of hydrophobins, the physiochemical nature of the hydrophilic side of the assembled form can be altered and a variety of new functionalities introduced. The fact that hydrophobins self-assemble at any hydrophilic-hydrophobic interface, irrespective of the chemical nature of the surface, therefore provides a generic approach to modify surfaces and make them interesting candidates for the use in various technical and medical applications.

Coating with genetic engineered hydrophobin promotes growth of fibroblasts on a hydrophobic solid.

Class I Hydrophobins self-assemble at hydrophilic-hydrophobic interfaces into a highly insoluble amphipathic film. Upon self-assembly of these fungal proteins hydrophobic solids turn hydrophilic, while hydrophilic materials can be made hydrophobic. Hydrophobins thus change the nature of a surface. This property makes them interesting candidates to improve physio- and physico-chemical properties of implant surfaces. We here show that growth of fibroblasts on Teflon can be improved by coating the solid with genetically engineered SC3 hydrophobin. Either deleting a stretch of 25 amino acids at the N-terminus of the mature hydrophobin (TrSC3) or fusing the RGD peptide to this end (RGD-SC3) improved growth of fibroblasts on the solid surface. In addition, we have shown that assembled SC3 and TrSC3 are not toxic when added to the medium of a cell culture of fibroblasts in amounts up to 125 microg ml(-1).

Biocompatibility of a novel tissue connector for fixation of tracheostoma valves and shunt valves.

Rehabilitation after laryngectomy often includes the use of a shunt valve and a tracheostoma valve to restore voice. To improve the fixation method of these valves, a new tissue connector has been developed, basically consisting of a ring that will be integrated into surrounding tracheal soft tissue. The valves can be placed in the ring. To test the principle of the tissue connector, a prototype consisting of a subcutaneous polypropylene mesh and a percutaneous titanium stylus was implanted into the backskin of 10 rats by a two-stage surgical procedure. We reasoned that if a firm connection can be realized with the skin, a firm connection with the trachea will also be possible. The subcutaneous part was implanted first, followed by the percutaneous part after 6 weeks. The complete tissue connector with surrounding tissue was removed 8 weeks later and examined histologically. The principle of the new tissue connector proved to be effective: hardly any epithelial downgrowth appeared, and adhesion of soft tissue was demonstrated. No infection or severe inflammation reaction was detected. The tissue connector seems appropriate for its intended use.

Characterization of human larynx carcinoma cell lines HLaC'79 and HLaC'82: a common origin but diverged malignancies.

As part of a study on the relationship of tumour phenotype and behaviour, we have characterized two head and neck squamous cell carcinoma cell lines, derived from human laryngeal carcinomas and designated HLaC'79 and HLaC'82. Cytogenetic analysis revealed that HLaC'79 and HLaC'82 shared 10 major chromosome rearrangements indicating that the cell lines had a common origin. In the extremely complex chromosomal patterns, abnormalities were found in chromosomes 1, 3 (surplus 3q) and 5 (i(5p) x 2). Both cell lines displayed constitutive expression of vimentin and were capable of anchorage-independent growth in agarose gels. However, in spite of their common origin specific differences were found. Cells of HLaC'79 were spindle shaped and formed tumours in athymic mice. In contrast, cells of HLaC'82 had a compact morphology, contained less vimentin, were more contact inhibited and were not tumorigenic. These results indicate that malignant transformation in HLaC'82 was partially reversed.

6-mercaptopurine: high-dose 24-h infusions in goats.

In vitro investigations have indicated the need for both prolonged exposure to 6-mercaptopurine (6MP) and the use of high concentrations to achieve maximal cell kill. After the customary oral administration the bioavailability of 6MP appeared to be low, and i.v. bolus injections resulted in short-lived high concentrations of 6MP, so prolonged infusions seemed rational. To test the feasibility of this approach 24-h infusions were given to goats. We used our improved HPLC method to quantitate 6MP and 6MP riboside (6MPR) in plasma, CSF, and urine. The concentrations of 6MPR were in excess of those of 6MP. Since 6MPR can easily be converted to 6MP, 6MPR acts as a depot for 6MP. Penetration of both 6MP and 6MPR into CSF was excellent. Of the total dose administered, 38% to 68% could be accounted for in the urine, with about equal amounts of 6MP and 6MPR. At doses of 20 and 10 mg kg-1 h-1 total concentrations of 6MP and 6MPR in excess of 100 microM were reached during 24-h infusions. However, all three experimental animals died due to toxicity. A dose of 2 mg kg-1 h-1 was tolerated; the total steady state concentration of 6MP and 6MPR in two experiments was about 10 microM. We conclude that the prolonged infusion of 6MP is feasible, and in view of the excellent penetration of 6MP and 6MPR into CSF, studies using prolonged infusions of thiopurines are warranted in man.

6-Thioguanine: high-dose 2-H infusions in goats.

6-Thioguanine (6TG) is poorly absorbed after oral administration. Bolus injections of 6TG result in high peak concentrations with relatively short-lived plasma concentrations. In vitro studies have shown the importance of prolonged exposure to 6TG. Therefore we administered 6TG by infusion at a dose rate of 2 mg/h over 2 h. In three goats we determined the plasma concentration-time curves of 6TG and its riboside (6TGR). A steady state was reached for 6TG and was almost reached for 6TGR within the 2 h of infusion. In one experiment we obtained several samples of CSF and observed good penetration of 6TG and 6TGR into CSF. Urinary excretion of 6TG and 6TGR was also quantitated. The amount of drug and metabolite excreted later than 4 h after the end of the infusion was negligible. By infusing 6TG, the problems of both erratic absorption after oral administration and acute renal toxicity after bolus injection, can be averted. In our opinion prolonged infusions of 6TG may be of advantage in humans suffering from actively proliferating malignant diseases, and thus should be studied.