I had never heard the term before today. From what I can tell, it's using antibodies to purify a cell population of interest. I would appreciate more details, especially in how it differs from "immunoprecipitation" of cells via cell surface markers, and FACS.
Immunopanning is essentially an immunoprecipitation (IP) of cells using an antibody immobilized to a solid surface, like a cell culture plate. Conventionally, an IP is performed using small agarose or magnetic beads (~50 to 150μm in size) conjugated to an antibody or Protein A/G, and can pull down individual proteins, protein complexes, and/or nucleic acid complexes.
Cells are much larger than protein complexes, obviously, and for the most part much smaller than beads, but they can be damaged during a traditional IP, so immunopanning utilizes a single solid surface coated with the specific antibody or cell-surface protein ligand of choice. In this article (free PDF here), retrograde axonal transport was used to selectively label specific populations of neuronal cells with the extracellular portion of cholera toxin β (CTB) adsorbed to fluorescent beads, then the cells were purified using an immobilized anti-CTB antibody.
Fluorescence-Activated Cell Sorting or FACS is another way to purify cell populations. Cells are labeled with a fluorescent marker (either conjugated to an antibody specific to an extra-cellular epitope, via absorption as in the example above, or by expression of a fluorescent transgene like GFP), then scanned in a flow cytometer. Cells which match certain parameters or "gates" in the machine's software are then diverted into a separate collection chamber, allowing for the generation of a quite pure cell population.
However, as the authors point out in their paper above, FACS requires an expensive piece of equipment (can be over US$100,000), can be harsh on the cells, and is dependent on the availability of an antibody to a cell surface marker specific to the desired population of cells. Immunopanning is much less expensive, is more gentle on the cells, and in the authors' case was really the only choice for purifying their cell population of interest, as no specific biomarkers were available. It certainly won't replace FACS in the long run, but is another tool for investigators to use when studying specific cell types.