We read with great interest the contribution by Drs Jakobiec and Bhat, entitled “Retrocorneal membranes: A comparative immunohistochemical analysis of keratocytic, endothelial, and epithelial origins.” The authors implemented an outstanding study and presented a good report. However, the article contains an error in Table 1 that may misguide the readers. In footnotes e and f , the authors have incorrectly footnoted and stated that the monoclonal antibody (mAb) CK CAM5.2 (Becton Dickinson, San Jose, California, USA) recognizes CK8 (53 kDa) and CK18 (45 kDa) exclusively. Obviously, anti-cytokeratin CAM5.2 has been inappropriately regarded as mAb specific for CK8 and CK18. Previous results by Makin and associates are attributed to breakdown products of CK8, giving smaller-molecular-weight fractions on immunoblots. In fact, BD Biosciences has globally marketed the anti-cytokeratin CAM5.2 (clone CAM5.2); and anti-cytokeratin CAM5.2 has a primary reactivity with CK8, but not CK18.

Additionally, the authors have stated in Table 1, footnote g , that the CK cocktail means a combination of CAM5.2 and AE1/AE3 (Signet Laboratories, Dedham, Massachusetts, USA), which provides a broad spectrum of both high- and low-molecular-weight cytokeratins. However, we would like to update the information about AE1/AE3, used in this study. Signet was originally created in 1989 as a commercial spin-off of Johnson & Johnson’s Cambridge Research Laboratories, Inc (CRL; Cambridge, Massachusetts, USA). In 2006 Covance acquired the assets of Signet Laboratories, located in Dedham, Massachusetts. Today, Covance (Emeryville, California, USA) markets the cytokeratin (AE1/AE3) monoclonal antibody for research use only (RUO) and the pan-cytokeratin (AE1/AE3) monoclonal antibody, purified (SIGNET), for in vitro diagnostics (IVD).

As stated in the supplier’s data sheets, antibody AE1 immunoreacts with an antigenic determinant present on most of the subfamily A cytokeratins, including cytokeratins with Moll’s designation 10, 13, 14, 15, 16, and 19 (molecular weights of 56.5, 54, 50, 50′, 48, and 40 kDa, respectively), but not on nos. 12, 17, and 18 (55, 47, and 45 kDa); whereas antibody AE3 reacts with an antigenic determinant shared by the subfamily B cytokeratins, including nos. 1 and 2, 3, 4, 5, 6, 7 and 8 (MWs of 65, 67, 64, 59, 58, 56, 54, and 52 kDa, respectively). In addition, anti-cytokeratin CAM5.2 reacts against human cytokeratin intermediate filament proteins of 48 kDa and 52 kDa, which were identified as cytokeratins 7 and 8, respectively.

As a result, both AE1/AE3 and CAM5.2/AE1/AE3 mAb cocktails directed against the same broad spectrum of both high- and low-molecular-weight cytokeratins with types 1-8, 10, 13-16, and 19. We believe this information is of great help to clarify the critical point that anti-cytokeratin CAM5.2 recognizes CK8, but not CK18. In addition, anti-cytokeratin CAM5.2 gives no advantage to the CK AE1/AE3 mAb.