PVRL4 (Nectin 4) was demonstrated to be an epithelial receptor for MV. This protein is a member of the poliovirus receptor-like proteins (PVRLs) that are adhesion receptors of the immunoglobulin superfamily 
. It is a 510-amino acid transmembrane protein with a predicted molecular mass of 55.5 kDa which migrates with a mass of 66 kDa on SDS polyacrylamide gels due to N-glycosylation. PVRL4 is an embryonic protein which has recently been shown to be a tumor cell marker for lung, breast, and ovarian adenocarcinomas 
. Like other nectins, PVRL4 is normally localized to the adherins junctions together with the cadherins. PVRL4 interacts with itself and the V domain of PVRL1, but not with other members of the same molecular family. Its cytoplasmic tail associates with the intracellular actin-binding protein, afadin 
. In humans it is expressed abundantly in the placenta and weakly in the trachea. However, in the adult mouse, PVRL4 transcripts were also found in the brain, lung, and testis 
. More recently, The Human Protein Atlas Project (www.proteinatlas.org
) has reported that PVRL4 is expressed abundantly in placental trophoblasts, glandular cells of the stomach, and adenocarcinomas of the lung, breast, and ovary. Moderate amounts of this protein appear to be expressed in the epithelial cells of tonsils, oral mucosa, esophagus, and the columnar epithelial cells of the nasopharynx and trachea. Smaller amounts are expressed in the lung macrophages and neuronal cells of the cerebral cortex.
The nectins can also function as entry receptors for several other viruses. PVR (CD155) is the prototype member of the family and was originally shown to be the receptor for poliovirus 
. PVRL1 (Nectin 1) serves as an entry receptor for herpes simplex virus (HSV). It is the major HSV receptor and mediates entry of all HSV-1 and HSV-2 strains as well as animal alphaherpesviruses 
. PVRL2 (Nectin 2) can also function as an entry factor for some herpesviruses 
. Alternatively, HSV can use another receptor of the TNF family called herpes virus entry molecule (HVEM). The fact that HSV-1 and HSV-2 use multiple receptors for cell entry appears to enable the virus to enter different cell types 
. This also appears to be the case with MV. The expression of Nectin 1 on the cell surface of tumors is also a predictor of oncolytic sensitivity to HSV in potential cancer therapy 
. Tight junction proteins can also serve as receptors for some viruses. For example, occludin is a major component of tight junctions that is used by hepatitis C virus 
and group B coxsackievirus as an entry factor 
. However, occludin shRNA experiments in MCF7 cells had no effect upon MV infectivity (data not shown).
The actual role of epithelial cells in MV infections has been controversial in light of recent experiments with macaques and SLAM transgenic mice. Alveolar macrophages, dendritic cells, and activated lymphocytes have recently been reported to be primary targets for IC323-EGFP wtMV and rMVKS-
EGFP strains, and there was only limited infection of epithelial cells in the studies with macaques over 5 to 9 days infection 
. Although lymphocytes and dendritic cells may be the major target and reservoir, virus has also been reported in the squamous stratified epithelium of the tongue and buccal mucosa and in the ciliated epithelium of the trachea at the peak of infection 
. It has been proposed that immune cells transmit MV to airway epithelial cells via a receptor on their basolateral surface and that these infected cells release virus from their apical cell surface to further infection and spread 
. It is possible that infected lymphocytes, expressing H and F proteins on their surface, attach to PVRL4 expressing epithelial cells to facilitate cell to cell spread of MV and syncytia formation late in infection. Again infected epithelial cells appeared at later times of infection than infected lymphocytes and dendritic cells 
. When a mutant MV that was blind to the epithelial receptor was used to inoculate macaques, clinical symptoms of measles were observed, but no virus was shed into the airways of these animals 
. This observation has led the researchers to conclude that the infection of epithelial cells occurred later in the disease, and was important for aerosol transmission of the virus.
We were initially surprised to observe that wtMV could infect adenocarcinoma cell lines from the apical surface in our Transwell filter assays. However, this finding may not be unreasonable since we were working with cancer cells where PVRL4 is highly upregulated, and it is expressed on both apical and basolateral cell surfaces. We have carefully considered the possibility that MV infection of adenocarcinoma cell lines using PVRL4 may not be totally relevant to understanding viral pathogenesis in the airways of the normal host. It has previously been reported that MV preferentially infects differentiated primary epithelial cells via the basolateral route, which is consistent with the location of PVRL4 at the adherens junctions of normal cells 
. Foci of infected cells derived from basolateral infections of primary epithelial cells did not appear to fuse and produce syncytia. This also warrants explanation. In addition, we and others were able to convert primary human SAEC grown in serum free to a MV susceptible phenotype by culturing them in 2% fetal calf serum. Flow cytometry indicated that little PVRL4 was produced on SAEC grown in serum free media, but the nectin was induced following transfer to serum containing media. PVRL4 may be expressed at higher levels on the apical surface during phases of rapid growth. In support of this, PVRL4 is highly expressed during embryogenesis 
. Upon revisiting the literature, we discovered that laboratories looking at basolateral infections of polarized primary human epithelial cell monolayers were waiting from 3–7 days post infection in order to see infected cell foci, and that apical infection of susceptible cancer cell lines was more efficient. This was consistent with our observations in and . In our hands, both MCF7 and NCI-H358 cells produced syncytia when infected via the basolateral route, although apical infection was much more efficient. Similar results were reported by Tahara et al
. who showed that polarized CaCo2 and HT-29 cells were preferentially infected via the apical surface 
. They also commented in their results section that the virus infected the basolateral side of the polarized monolayer much less efficiently than the apical side. Other investigators also obtained similar results using a laboratory strain of MV to infect CaCo2 cells 
. However, when they infected primary human airway epithelia via the basolateral surface, they saw more virus replication compared to infection of the apical side. Mutations in the putative epithelial receptor binding region of MV H protein, to produce an epithelial receptor blind virus, blocked basolateral infection of primary cells as well as apical infection of lung adenocarcinoma cell lines, indicating that the same receptor is probably used in each case 
Other laboratories have shown that PVRL4 (Nectin 4) is up-regulated on breast, lung, and ovarian cancer tumors and cell lines 
. We extended this observation to colon carcinoma cell lines including DLD-1, HT29, and LoVo cells. There are anecdotal reports in the literature where natural MV infections were shown to reverse cases of Burkitt's lymphoma and Hodgkin's disease 
. Given that these tumors express SLAM/CD150, one can presume that wtMV infected the tumors and triggered immune attack against them. Vaccine strains of MV have previously been engineered to recognize cancer cells by artificially manipulating the H receptor binding protein 
. Clinical trials are currently in progress at the Mayo Clinic (Rochester, MN) for ovarian cancer, pancreatic cancer, glioblastoma, medulloblastoma, and multiple myeloma 
. It may be possible that a natural tropism of MV for lung, breast, colon, bladder, and ovarian adenocarcinomas can also be exploited for future oncolytic therapies. The use of SLAM/CD46 blind MV that retained the ability to bind PVRL4 could constitute a potential therapeutic vaccine against adenocarcinoma. Safety issues, background infection, and the effect of pre-existing antibodies from MMR vaccination will obviously have to be addressed before human clinical trials could even be considered.
This study identifies PVRL4 (Nectin 4) as an epithelial receptor for MV. PVRL4 is expressed at low to moderate levels in normal tissues but is highly up-regulated on the surfaces of adenocarcinoma cells. The interaction is highly specific, since MV does not recognize other members of the PVRL family and prefers the human receptor over homologues in the mouse. Further experiments with differentiated primary epithelial cells in culture and the use of human epithelial explants will be required to validate the role of PVRL4 in infections of normal epithelial cells and establish its importance in measles pathogenesis. These studies are currently underway in our laboratory.