As described in our title, “Internalization of cholera toxin by different endocytic mechanisms”, we studied the endocytic pathways by which cholera toxin (CT) can be taken in from the cell surface. We have previously studied CT transport through the Golgi apparatus and to the ER, as well as cAMP production (Sandvig et al.,1996), and we were among the first to demonstrate this transport step for CT. Similarly, as first shown for Shiga toxin(Sandvig et al., 1992), other toxins are transported retrogradely before entry into the cytosol (for a review, see Sandvig and van Deurs,2002).
However, studies of CT entry are of interest not only because this toxin can increase the level of cAMP in some cells and cause diarrhoea, but also because CT has been commonly used to label GM1 and to study endocytosis from caveolae. In our study (Torgersen et al.,2001), we used three different model systems to modulate uptake by endocytic pathways: (1) Caco cells transfected with caveolin to create caveolae at the cell surface; (2) Hela cells with inducible synthesis of mutant dynamin, which have been reported to inhibit pinching-off of vesicles from both clathrin-coated pits and caveolae; and (3) BHK cells with inducible synthesis of antisense-clathrin heavy chain (cells in which clathrin-dependent endocytosis can be shut off selectively). The data obtained from the three systems reveal that both clathrin-dependent and clathrin- and caveolae-independent mechanisms can lead to endocytosis of CT. This study does not allow us to conclude whether caveolae can be responsible for endocytosis of cholera. Although caveolae (with caveolin) are quite stable structures in some cell types (Thomsen et al.,2002), the toxin itself might affect the stability. To date, no publications have addressed this question. However, our study(Torgersen et al., 2001)clearly shows that different endocytic pathways can be involved. This is in agreement with recent data published by other laboratories(Nichols et al., 2001; Shogomori and Futerman,2001,Shogomori and Futerman,2001). We did not make any attempt to answer whether toxin taken in by the various endocytic mechanisms can elicit a biological response(Torgersen et al., 2001). This is of course an important question but was not addressed in our study.
When one investigates the effect of a certain drug (which quite often has more than one effect on cells) or, for instance, the importance of cholesterol(either by adding drugs, removing cholesterol or adding cholesterol) on the action of a toxin, a reduced or increased effect can be caused by an effect on the endocytic uptake, or by an effect on a later step, such as endosome to Golgi transport of the toxin. Along these lines, it has recently been published that transport of CT (Shogomori and Futerman, 2001,Shogomori and Futerman, 2001), ricin(Grimmer et al., 2000) and the Shiga toxin B subunit (Falguieres et al.,2001) from endosomes to the Golgi apparatus are affected by changes in cholesterol. Also, when comparing the effect of a drug on the cytoplasmic action of various toxins, there may not necessarily be a difference because of the different endocytic mechanisms used by the toxins,but a different response could be caused by different intracellular pathways(cholera/diphtheria toxin). Alternatively, the drug could have a direct effect on the target molecule, for instance on the activity of a membrane-associated target such as adenylyl cyclase, which in itself could be regulated by, for instance, cholesterol or drugs affecting cholesterol. Studies with proteoliposomes have even shown that coupling between Gs and adenylyl cyclase can be dependent on the cholesterol:phospolipid ratio(Bai and Youguo, 1998). In order to avoid such complications in our endocytosis studies, we concentrated on the endocytic uptake by directly measuring the internalization from the cell surface (Torgersen et al.,2001).
It should be noted that there are cell-specific differences (as discussed in our article) when it comes to uptake from the cell surface (as well as to intracellular routing of toxins). Thus, whether a toxin is associated with lipid rafts (Falguieres et al.,2001), and to what extent it is transported to the Golgi apparatus, is clearly cell-type dependent and can be dependent on the type of fatty acid in the toxin receptor(Falguieres et al., 2001; Sandvig and van Deurs, 2002; Lingwood, 1999). That endocytosis of CT can occur independently of filipin addition is supported by a previous study (Shogomori and Futerman,2001,Shogomori and Futerman,2001). In fact, when clathrin-dependent endocytosis is reduced by antisense-clathrin induction, or when dominant-negative mutant dynamin is induced to inhibit both clathrin- and caveolae-dependent endocytosis, there is no further decrease in the uptake of CT by extraction of cholesterol with mβCD (M. L. Torgersen, B.v.D. and K.S., unpublished). Thus, in the cells we have studied, CT can be endocytosed even under such conditions.
In conclusion, it is clear that CT can be taken in by various endocytic mechanisms, and that more has to be done to characterize these mechanisms as well as the intracellular transport of CT. To fully characterize the transport of CT, investigation of each step will be necessary.