Hematopoietic Colony Forming Unit: Development of a High-Throughput CFU Assay Strategy by the Use of High-Resolution Digital Images Stored in a Laboratory Information System.

Background: Umbilical Cord Blood (CB) is an increasingly accepted graft source for patients lacking related donors. Graft characteristics currently used as determinants of quality and engraftment potential of CB units include the enumeration of Total Nucleated Cell (TNC), CD34+ and Colony Forming Units (CFU). Among those, the 14 day-CFU assay is the only one that determines the functional state as well as the number of hematopoietic progenitor cells. Thus, CFU in pre and post cryopreservation/thawing specimens are strong independent predictors of CB graft engraftment. Traditionally, evaluation of CFU growth is performed by light microscopy (traditional classification), is time consuming, subjective and difficult to standardize.

Aims: A) To evaluate whether the incorporation of high-resolution digital imaging and colony staining could make reading the traditional CFU assay objective and readily standardizable. B) To validate the new strategy vs traditional classification. C) To develop a computer based laboratory information management system (LIMS) to support high-throughput CFU assay.

Methods and Results: After 14 days of CB culture (CFU assay-Stem Cell Technologies), an image of the 35 mm culture dishes was captured using a high-resolution photographic camera based digital imaging system, which achieves a resolution of 7.6 μM per pixel and thus, allows a clear view of all colonies in the dish with their barcoded IDs. A short one-step staining with MTT (3-[4,5-dimethylthiazol-2yl]-2,5-diphenyltetrazolium bromide) allows an even better definition of CFU-GM/E; CFU-GM and CFU-E by bestowing a specific color on each type (dark purple, purple and red respectively) against a uniformly clear background. A good correlation was observed after comparison of the new strategy against traditional enumeration (R² linear= 0.95; n= 122 culture dishes evaluated). Low variation was observed after 151 cultures were independently classified and enumerated by three different operators (CV%= 8.9; range 1–27%). Sample plating introduced variation of the CFU assay, in an experiment where nine CB samples were evaluated by multiple plating (Intra-assay CV%= 21.9 %; range 3.4–34.5% and Inter-assay CV%= 23.3%; range 12.6–35%). A computer based laboratory information (LIMS) was developed to store all culture dishes, linked by unique barcoded ID labels to a specific CB unit, and including CB image, incubator location, plating and counting dates, as well as detailed colony enumeration. This system has been used for 5 months in our laboratory and more than 3,000 CB units have been tested in duplicate (average: 30 CB/day). Images of an average of 60 culture dishes and MTT staining can be performed in less than two hours. The specific coloration of CFU colonies allows faster classification and enumeration and thus, permits a more precise analysis of CFU colonies and its relation with CD34+ cell content and post-transplant engraftment.

Summary: With this new strategy, CFU can be objectively visualized, differentiated and counted; the digital images can be stored for future review and refined classification. The described system provides computerized information on optical assay parameters and has become an invaluable tool supporting high-throughput implementation. Thus, the combination of high resolution imaging, one-step staining and the traditional CFU assay overcome most technical challenges of the conventional method, supporting standardization and yielding high reproducibility to the assay which can be easily implemented in CB banks where large numbers of samples need to be tested daily.