Previous studies on the assembly of progesterone receptor (PR) complexes in vitro have suggested that PR assembly is a dynamic, ordered process involving at least eight nonreceptor proteins. One of these proteins, p60, appears transiently during assembly and is not a component of functionally mature PR complexes. In the present study we observe that a monoclonal antibody specific for p60 can, on the one hand, inhibit formation of mature PR complexes containing heat shock protein 90 (hsp90), p23, and immunophilins and, on the other, enhance recovery of early PR complexes containing hsp70 and Hip (p48). This observation supports a model in which p60 functions at an intermediate stage of PR assembly to facilitate formation of subsequent PR complexes lacking p60. Since p60 is typically found in a complex with hsp90 and hsp70, we have further characterized its interactions with these proteins. P60 can bind either hsp70 or hsp90 independently and in an ATP-independent manner. Since hsp90 and hsp70 do not readily associate on their own, it appears that p60 is the central organizing component of an hsp90-p60-hsp70 complex. Mutational analysis of p60 indicates that the N terminus is required for hsp70 binding, and a central region containing tetratricopeptide repeat motifs is necessary for binding hsp90 and hsp70. The hsp90-p60-hsp70 multichaperone complex is highly dynamic and does not appear to be affected by the hsp90-binding drug geldanamycin. The interactions of hsp70 and hsp90 in intermediate PR complexes are shown to be distinct from their separate interactions in early PR complexes (hsp70) or in mature PR complexes (hsp90). From these results, it appears that p60 is a key mediator in the chaperoned assembly and functional maturation of PR complexes.

A cDNA for human FKBP51 has been cloned and sequenced, and protein products have been expressed in both in vitro and bacterial systems. The deduced amino acid sequence for human FKBP51 is 90% identical to sequences of recently described murine proteins and is 55% identical to the sequence of human FKBP52. Human FKBP51 mRNA is expressed in a wide range of tissues, and the protein has peptidylprolyl isomerase activity that is inhibited by FK506 but not cyclosporine. FKBP51 is the same as a previously described progesterone receptor-associated immunophilin that, similar to FKBP52 and cyclophilin 40, is an Hsp90-binding protein and appears in functionally mature steroid receptor complexes along with Hsp90 and p23. Each of the three receptor-associated immunophilins displays interactions with progesterone receptor that are more dynamic than Hsp90-receptor interactions. Whereas FKBP52 and FKBP51 compete about equally well for binding to Hsp90 in a purified system, FKBP51 accumulates preferentially in progesterone receptor complexes assembled in a cell-free system. This observation provides a precedent for differential interactions between Hsp90-associated immunophilins and target proteins such as steroid receptors.

The hsp70-interacting protein Hip participates in the assembly pathway for progesterone receptor complexes. During assembly, Hip appears at early assembly stages in a transient manner that parallels hsp70 interactions. In this study, a cDNA for human Hip was used to develop various mutant Hip forms in the initial mapping of functions to particular Hip structural elements. Hip regions targeted for deletion and/or truncation included the C-terminal region (which has some limited homology with Saccharomyces cerevisiae Sti1 and its vertebrate homolog p60), a glycine-glycine-methionine-proline (GGMP) tandem repeat, and a tetratricopeptide repeat (TPR). Binding of Hip to hsp70's ATPase domain was lost with deletions from the TPR and from an adjoining highly charged region; correspondingly, these Hip mutant forms were not recovered in receptor complexes. Truncation of Hip's Sti1-related C terminus resulted in Hip binding to hsp70 in a manner suggestive of a misfolded peptide substrate; this hsp70 binding was localized to the GGMP tandem repeat. Mutants lacking either the C terminus or the GGMP tandem repeat were still recovered in receptor complexes. Truncations from Hip's N terminus resulted in an apparent loss of Hip homo-oligomerization, but these mutants retained association with hsp70 and were recovered in receptor complexes. This mutational analysis indicates that Hip's TPR is required for binding of Hip with hsp70's ATPase domain. In addition, some data suggest that hsp70's peptide-binding domain may alternately or concomitantly bind to Hip's GGMP repeat in a manner regulated by Sti1-related sequences.

Purpose: Binary outcomes (toxicity=1, no‐toxicity=0) are frequently used as inputs to fit parametric models. However, in reality, outcomes span the range in‐between these extremes; binary values are used for lack of more precise quantification. We propose a method to estimate non‐binary outcomes, which can then be used to more accurately fit conventional parametric models (e.g., Lyman‐probit, relative‐seriality). Methods and Materials: The non‐binary outcomes were estimated using 4 machine learning algorithms. The algorithms used variables selected from the available data to best‐fit binary outcomes. Based on the fitting, each algorithm outputted binary estimates of whether or not the patient suffered toxicity. By fitting each algorithm to randomly selected subsets of the patient data, multiple binary toxicity estimates were generated for each patient. Averaging these estimates for each algorithm produced a non‐binary estimate. Furthermore, averaging these non‐binary estimates over all algorithms reduced prediction bias. This final algorithm‐averaged non‐binary patient estimate was then used as input to conventional parametric algorithms. This method is demonstrated in the context of radiotherapy‐induced pneumonitis. Results: One‐hundred estimates from each algorithm × 4 algorithms were averaged to produce a non‐binary toxicity estimate for each patient. Both the original binary toxicity outcome and the non‐binary estimate were fitted to the parametric Lyman‐probit and relative‐seriality models. Both parametric models had large confidence interval limits when fitted to the original binary outcomes. The fits improved when using the non‐binary estimates. The improvement was substantially better for the Lyman‐probit model, but only marginally so for the relative‐seriality model. Thus, the Lyman‐probit model is better suited to modeling radiation‐induced pneumonitis, a conclusion that would not have been possible from the original binary outcomes. Conclusions: Non‐binary estimates from machine learning algorithms can be used to not only generate better fits to conventional parametric models, but also to deduce which model is methodologically better suited.

Purpose: Stereotactic body radiation therapy (SBRT) requires high precision of patient position and target localization. For SBRT patients positioning, cone beam CT imaging has been widely used, generally with zero couch rotation. The purpose of this study is to implement Stereotactic radiation surgery (SRS) patient positioning technology to SBRT by expanding patient positioning with couch rotation. Methods: Currently we are using Varian® Novlis Tx for treating SBRT patients implementing CBCT. BrainLAB® X‐ray imaging system in conjunction with optical guidance is primarily used for SRS patients. CBCT and X‐ray imaging system were independently calibrated with 1.0 mm accuracy. For daily imaging QA, we set up a Penta™ imaging phantom with infrared markers. The imaging phantom has two image isocenters for CBCT and X‐ray imaging respectively with CBCT isocenter offset from the X‐ray imaging isocenter. The X‐ray imaging system was implemented through BrainLAB® ExacTrac system with CBCT localized position at the initial zero position for the X‐ray imaging system. For the other couch positions, X‐ray images were fused with patient DRRs for positioning. Results: For daily imaging QA, the longitudinal, vertical and lateral coordination between CBCT and X‐ray imaging average 0.4+/−0.6, 0.1+/−0.6 and 0.6+/−0.7 mm. The shift from the CBCT imaging isocenter to the X‐ray imaging isocenter is 0.7+/−0.5 mm accuracy for a 6‐month period of tracking.Patient position accuracy: After initially localizing the patient with CBCT at the zero couch position, we then position the patient with the X‐ray imaging system. The computed translational and rotational shift accuracy are 0.6+/− 0.6 mm and 0.5+/−0.3 degree respectively, based on 39 SBRT patients couch rotations. Conclusion: Accurate coordination of CBCT and X‐ray imaging in conjunction with optical imaging guidance can be expanded to patient positioning with couch rotation. The X‐ray imaging capability at rotated‐couch positions improved the physician confidence level during SBRT treatment.