A HEK293 Cell Database

A HEK293 Cell Database- put together by Gerry Shaw

Introduction

    HEK 293 cells were generated by transformation of human embryonic kidney cell cultures (hence HEK) with sheared adenovirus 5 DNA, and were first described in 1977 (Graham et al., J. Gen Virol 1977 Jul;36(1):59-74). The story of this cell line is quite an interesting one, getting to the very roots of modern molecular biology. Frank Graham was working with Alex Van der Eb in the University of Leiden in Holland, and developed the calcium phosphate method of transfecting eukaryotic cells, an astonishing technique described in a classic paper (Graham and Van der Eb. Virology 1973 Apr;52(2):456-67); who would have thought that you would only have to mix DNA with a bit of calcium phosphate to get the DNA incorporated and even expressed in mammalian cells? The story of this discovery was described by Graham himself for Current Contents in 1988 (link to *.pdf here). By 1977 Graham had moved to Canada, and from there produced the other classical paper describing the derivation of the HEK 293 cells. There is also an interesting Current Contents article by Graham describing this work (link to *.pdf here). Also, check out the transcript of an FDA meeting entitled "Vaccines and related biological products advisory committee" which took place in May 2001. At this meeting Alex Van der Eb describes the work that led to the HEK 293 derivation. I was particularly fascinated that he actually speculated that these cells may have originated from a rare neuronal cell in the kidney cell cultures, since at about the same time but completely independently we had come to a similar conclusion (link to *.pdf here- Dr. Van der Eb's part starts on page 77, and the specific comment about the potential neuronal origin of 293 cells is on page 85).

    Presented here are a summary of the initial results obtained by DNA microarray analysis of 293 cells. This analysis was stimulated by the observation that 293 cells stain strongly and specifically with antibodies to several neurofilament proteins, which are generally thought of as excellent markers for neuronal lineage cells. As with quite a lot of interesting stories in Science this was as a result of a completely unexpected finding; I was making monoclonal antibodies to the beta-adrenergic receptor kinase 1 (beta-ARK), which is supposed to be endogenously heavily expressed in 293 cells. I was at the point of screening by immunocytochemistry on fixed 293 cells, which I had bumed from another lab. The staining pattern was likely to be weak and diffuse at this stage so I needed a good control. So I reached into the fridge for a monoclonal antibody which would have no business staining a kidney cell, and since one of the things I work on is neurofilaments, I got out one of our monoclonals to NF-M, one of the major neurofilament subunits. I was surprised to find that the 293 cells all stained strongly and in a clearly filamentous pattern with this antibody. Since I had made and characterized the particular antibody myself I knew immediately that this was no artifact, these cells really must contain NF-M. Further biochemical and immunological experiments verified the presence of large amounts of the major neurofilament subunits NF-M and NF-L, and smaller amounts of alpha-internexin. Small amounts of NF-H were detected in a few cells, so that the four major neurofilament subunits were detected in 293 cell cultures. We also detected vimentin and the two primitive keratins K8 and K18, both immunocytochemically and by immunoblotting. The keratins are primitive in that they are the first expressed in development, so are characteristic of the earliest differentiating tissues of the body. Early differentiating neurons typically express NF-L, NF-M and alpha-internexin, but only express NF-H in large amounts later, as seen in 293. We did not detect desmin, GFAP, peripherin or any the high molecular weight keratin subunits. The pattern of expression is quite similar to that seen in PC12 and Ntera-2 cells, which also express a mix of neurofilament subunits and the two primitive keratins. So the 293 cells were similar in several respects to early differentiating neurons and to two well defined neuronal lineage cell lines.

    Further studies revealed that several other cell lines derived from human embryonic kidney cultures by adenovirus transformation also express neurofilaments, as did adenovirus derived cell lines from human embryonic retina, and the basic pattern of neurofilament expression was similar to that of 293 cells. While adenovirus efficiently generate cell lines from human embryonic retinal cultures, it has always been very difficult to generate lines from human embryonic kidney cultures. We hypothesize that a rare neuronal lineage cell type is the target for adenovirus transformation in these kidney cultures, which would obviously explain why it is difficult to obtain these cells from embryonic kidney cultures but much easier from embryonic retinal cultures. Literature searches reveal that there are in fact a few neuronal lineage cells in developing mammalian kidney, which we propose are the preferred targets for adenovirus. Further literature searches suggest that adenoviruses tend to specifically transform neuronal lineage cells, in line with the hypothesis presented here. We conclude that 293 cells have a phenotype surprisingly similar to established neuronal lineage cell lines such as PC12 and Ntera-2 cells, and that adenoviruses target specifically neuronal lineage cells.

    Microarray analysis of 293 cells using Affymetrix and Clontech arrays revealed mRNA encoding many other proteins normally expressed in neuronal lineage cells, and the data is summarized here in Table 1. For access to one of the full data sets press here. Apologies to those of you who tried to access this in the last few months, there have been various problems associated with security measures on our local server- this link should work now, but you will have to select the "293.xls" file manually). The Affymetrix chips score in terms of "Average Difference", which is essentially the amount of hybridization detected on 16 different 25 mer specific probes specific for different regions of the mRNA to be tested. The hybridization signals obtained are compared to the signals from 16 different 25 mer control probes which are identical to the experimental probes except that they include a single deliberate base mismatch. The data is an Excel spreadsheet ordered by the average difference column, so that those mRNA producing the strongest signals are at the top.The highest average difference values were 100,000 and above, and correspond to ribosomal RNA,b-actin or other well known abundant RNA. The threshold of reliable detection is probably average difference values in the range of 500-1,000. In a few cases the Affymetrix chip has multiple probes for a particular mRNA. Where multiple probes are used the table reveals this by providing multiple average difference values, and also the "absolute call" value, which is "P" for present, "M" for maybe or "A" for absent. In some cases these results are in conflict and in these cases the data should be interpreted with caution. For specific questions email me at shaw@mbi.ufl.edu. Tables 2 - 7 proved a summary of data on mRNA encoding G protein coupled receptors, receptor tyrosine kinases and other important families of molecules. The mRNA indicated with a * in Table 1 were derived from Clontech cDNA arrays which do not use the average difference measurement. You can download a *.pdf of these findings as published in Faseb Journal Express here.

Table 1: mRNA detected in 293 cells normally expressed specifically in neuronal cells

Table 2: G Protein coupled receptors

Table 3: TGFbreceptor family

Table 4: Cytokine Receptors

Table 5: Apoptosis related

Table 6: Ligand gated ion channel molecules

Table 7: Insulin related molecules

Souce: Gerry Shaw    2016-07-18