Sirs

We recently had the opportunity to read the Opinion article by Matthew Albert1 that was published in Nature Reviews Immunology. Because this article is probably of interest to many scientists, we felt that it was important to bring attention to a factual error regarding one of our studies.

Albert contends that apoptosis, together with increased expression of peptide–heat-shock protein (HSP) complexes, is an important source of antigens for dendritic cells (DCs), and that it is superior to necrosis in this regard. On page 227, Albert refers to our study that was recently published in Blood2, in which we compared the capacity of apoptotic and necrotic vaccinia-virus-infected monocytes to load DCs with antigens. He states, “As further evidence for an active process being involved in their so-called necrotic death ... they report upregulation of HSP expression and argue that peptide–HSP complexes in their lysates are mediating antigen transfer.” However, in our paper2, we do not examine the expression of HSPs in infected cells — apoptotic, necrotic or otherwise — and consequently, we do not make the argument that HSPs are mediating the transfer. In fact, the main message and conclusion of our paper was that, regardless of the source of antigens (from apoptotic cells or so-called necrotic cells), a common pathway is used to present cell-derived antigens, and this depends on endosomal pre-processing, proteasomes and transporter for antigen processing (TAP). Furthermore, we strongly believe that the antigen delivered by apoptotic and necrotic cells for cross-presentation by DCs is in the form of protein or protein aggregates, and not peptides2. In vivo studies by Shen and Rock support this by elegantly showing that cellular proteins are the source of antigen for cross-priming in vivo3.

Several studies of cross-presentation show dependence on degradation of exogenous antigen by the proteasome, implying that the antigen is not delivered as peptides but as protein or large protein fragments that require degradation into peptides in DCs2,4,5. Albert cites his own unpublished data showing that pre-processing in apoptotic cells is sufficient to transfer antigens to DCs (page 227), but he fails to mention that our study2 does not support his argument.

Finally, Albert chooses to ignore several in vitro and in vivo studies (using animal models and humans) that have shown that lysates of primary cells, transformed cells or cell lines, which were not undergoing apoptosis, can deliver antigens to DCs and induce immunity6,7,8,9,10,11. Instead, he cites one study12 in support of his argument that apoptotic but not necrotic tumour-cell vaccines induce potent immune responses.

Albert is entitled to his opinion. However, because an uninformed reader will interpret his reference to our data as fact, we feel that it is essential that our comments are brought to the attention of readers.