This page was created as an undergraduate assignment for Davidson College.

Review Paper Project

Integrin a3b1 (CD 49c/29) Is a Cellular Receptor for Kaposi’s Sarcoma-Associated Herpesvirus (KSHV/HHV-8) Entry into the Target Cells

Shaw M. Akula, Naranatt P. Pramod, Fu-Zhang Wang, and Bala Chandran. 2002. Cell. 108: 407-419.

Reviewed by Danielle Hyun-jin Choi

The researchers of this article studied the role of integrins to act as receptors for the herpesvirus envelope glycoprotein B (gB). Human herpesvirus-8 (HHV-8) is associated with Kaposi’s Sarcoma (KS), most commonly seen in AIDS patients. The patients develop malignant skin tumor that results in multifocal purplish colored sores that eventually form nodules, small lumps of hard swelling tissues, even within the lungs and intestines.3

HHV-8 envelope protein contains the RGD motif with which integrins interact. To study the binding of HHV-8’s RGD to integrins, the researchers used artificial RGD peptides and antibodies against RGD-binding integrins to detect inhibition of HHV-8 infection. The artificial RGD protein constructs, the antibodies against RGD amino acid, and the soluble a3b1 integrin significantly blocked infection.
The researchers also reasoned that if more RGD-dependent integrins were expressed on the cell membranes, the more receptors there would be for HHV-8, and more infection would be obtained. And they did observe increased infection in Chinese hamster ovary cells, when human a3 integrin was artificially expressed on the cell surface.

Immunoprecipitaion method was used to study the binding of anti-gB antibody to specific chains of the integrin proteins. Further investigation involved the detection of integrin-dependent focal adhesion kinase (FAK) activation with soluble a3b1 integrins and anti-gB antibodies. Interestingly, the HHV-8 infection inducing factors seemed to phosphorylate FAK.

Fig 1. RGD peptides inhibit HHV-8 infection.
Green Fluorescent Protein (GFP-HHV-8) and antibodies against HHV-8 ORF73 were used to detect the level of infection. Control peptides did not include RGD amino acids, while the treatment peptides contained RGD. Panels A and D were treated with control peptides and were mock infected, so no signal of infection was detected. In panels B and E, the cells were treated with the control peptides, and were infected with GFP-HHV-8. As expected, I see signals for infection. The interesting data come in C and F: when the cells were treated with RGD peptides and with GFP-HHV-8, I see relatively less signal than shown in B and F. I agree with the authors that the RGD peptides inhibited GFP-HHV-8 infection. I also liked the idea of screening for the HHV-8 ORF73 proteins, because the only source of those proteins is HHV-8, and allowed the researchers to measure infection by non-GFP HHV-8, since the “GFP-HHV-8” actually contained both wild-type and recombinant viruses.

Fig 2. Antibodies against RGD peptides and other integrin ligands also inhibit HHV-8 infection.
The researchers hypothesized that if integrins are acting as receptors and HHV-8 envelope proteins as ligands, blocking the ligands should inhibit binding to the receptors. Also, if the receptors (integrins) have already bound to other ligands, the receptors are blocked, so that the virus envelope protein cannot bind to integrins. (See diagram below)
Indeed, the data shown here seem to support that hypothesis. Graph A shows a case for “blocking the receptor” by RGD peptides. Compared to the non-RGD peptide, higher level of inhibition was observed for RGD peptide treatment. Graph B shows a case for “coating the ligand” by antibodies against virus envelope proteins. Since antibodies against RGD peptide coated the ligands around the virus, higher inhibition was observed compared to non-specific antibodies. Graph C is another case for “blocking the receptor” mechanism of inhibition by known integrin ligands (fibronectin), compard to a variety of other ligands.

Fig 3. Soluble a3b1 integrin inhibits HHV-8 infection.
Now the researchers have moved on to identify which integrins are involved in HHV-8 gB binding and infection. They have used a variety of anti-integrin antibodies, for some a units and b units separately, and some in combinations. They measured % inhibition of GFP-HHV-8 infection, and found that a3 was the most effectively inhibited integrin, and then b1 and a2b1 integrins were similarly inhibited in two different types of cells (Graph A). Curiously, they did not include a3b1, which they suggested to be the receptor for HHV-8 gB. They show that inhibition by anti-integrin antibodies was dose-dependent (Graph B). I think one of the most important figures is Graph C, because the graph depicts the specificity of ligand-receptor interaction for HHV-8 gB and a3b1 integrin. See the diagram below for why soluble a3b1 integrin would inhibit infection.

Fig 4. Determining the relative abundance of a3b1 integrin.
Flow cytometric analysis was used. The higher the integrin expression level was, the more intense the fluorescence signal was detected for this FACS data. A is the control (no integrin detected), B for avb3 integrin expression, C for a3, and D for b1. The data suggests that significantly more a3 and b1 integrins were expressed and detected, compared to the control expression level. Table E shows the relative abundance of target integrin expression in different types of cells with Mean Fluorescence Intensity value in paranthesis. I note that the relative abundance of a3 and b1 is well over 90% for BJAB and HFF cells.

Fig 5. Expression of human a3 integrin in hamster ovary cells (CHO-B2) increase HHV-8 infection.
As the percentage of human a3 expression increased in CHO-B2 clone D5 cells (2) and even more in CHO-B2 clone B3 cells (3) compared to the control (1), the intensity of fluorescence increased, indicating an increase in GFP-HHV-8 infection in those cells. In Panel B, more signal for GFP-HHV-8 infection is observed from B3 cells (2 and 4), than D5 cells (1 and 3). This panel seems to be missing the control. The researchers should have shown us a GFP detection of cells that were not infected with GFP-HHV-8. Also, panel B seems redundant since the differences in fluoresence intensity for expression of GFP-HHV-8 in D5 virsus B3 cells were already shown in panel A (2 and 3). Panel C shows the increased susceptibility to HHV-8 infection as a function of expression of human a3 integrins. The middle bars represent inhibited infectivity due to the antibodies added, thus coating the virus envelope proteins with antibodies to inhibit infectivity, and this procedure proves that the integrins of interest and GFP-HHV-8 are the key players at work. Anti-b4 cannot coat the virus proteins, so no inhibition is observed (the right most bars). It seems fairly convincing that the increase in expression of human a3 integrins resulted in the increase in infectivity. But I think the mechanism by which the hamster b1 and human a3 integrins interact was not addressed clearly, which is critical for the explanation of increase in infectivity.

Fig 6. Immunoprecipitation of the virus a3 and b1 complexes with anti-gB antibodies.
The main point of this figure is to show that the anti-gB (HHV-8 glycoprotein B) antibodies specifically immunoprecipitated a 150/110 kDa heterodimer protein, which was identified as a3 and b1 (lanes 7, 8, 10, 11). However, anti-gB antibodies did not immunoprecipitate a1 chain of the integrin. Heparinase was added to prevent undesirable interaction between HHV-8 and the ubiquitous cell surface haparin sulate-like molecule. Their claim that HHV-8 probably recognizes a specific conformation of a3b1 only from this data seems bold to me. Further study seems to be required to see the exact conformational interaction between HHV-8 gB and the a3, b1 chains of the integrins. Panel B shows that only heparin blocked binding of HHV-8 to HFF cells, while other RGD peptides, antibodies to RGD gB and other a3b1 associated proteins did not. The failure to inhibit binding suggests that a3b1 integrin interacts with HHV-8 at a postattachment step of infection. Panel B study seems too open-ended to me.

Fig 7. Panel A shows the change in distributions of Focal Adhesion Kinase in the cell after HHV-8 infection or after lysophosphatidic acid (LPA) treatment. For FAK to change distributions by colocalizing with vinculin, FAK must be phosphorylated first. After 5 minutes of HHV-8 infection, FAK showed colocalization with vinculin in panel A. It seems convincing enough that FAK was phosphorylated due to HHV-8 infection and changed distributions. For panel B, the gradual increase in the intensity of the bands from lane 3 to 5, suggest that longer exposure to HHV-8 induced more phosphorylation of FAK. (The bands indicate that anti-phospho-FAK antibodies bound to the phosphorylated form of FAK.) For C, soluble a3b1 integrins were added in decreasing amount from lanes 2 to 4, and the increase in the band intensities sugguest that the soluble integrins inhibited phosphorylation of FAK. For D, anti-gB antibodies were added in decreasing amount from lanes 2 to 4, and the increase in the band intensities suggest that the anti-gB blocked phosphorylation of FAK. These data seem fairly convincing and clear to me that a3b1 integrins play a direct role in phosphorylation of FAK.

Overall, the authors of this study have indeed convinced me that a3b1 integrins are acting as receptors to HHV-8 glycoprotein B. Their hypothesis that a3b1 integrins are receptors that bind to HHV-8 gB to permit entry of the virus into the host cells seems congruent with the data they have given. There are some loose strings I have noted in the figures explanations, but this paper is fairly straight forward.
The identification of a3b1 integrins as receptors to HHV-8 opens up many future study directions. First of all, since present study was done in vitro, only in test tubes, it would be sensible to study the in vivo roles of these receptors in HHV-8. Since GFP expression does not destroy the cells or inturrupt with the cell functions (at least there have not been evidence that GFP inturrupts with the life of the cells), we will be able to detect precisely which cells in the organism have been infected by HHV-8 with GFP constructs, when the infection occurs at what periods of time in the organism’s life, and relative abundance of expression in each cell type. The researchers have already mentioned that HHV-8 infection have been detected in vivo in human B cells, endothelial cells, epithelial cells and others in the introduction. So now, with the knowledge such as that soluble a3b1 integrins can block HHV-8 infection, that a3b1 integrin ligands and anti-RGD antibodies can also inhibit infection, we can now investigate how effective the inhibition is in vivo. This type of study may lead to discovery of vaccines that are effective in preventing or treating HHV-8 infections. Since HHV-8 has been known to be associated with Kaposi’s Sarcoma, such vaccine may be revolutionary in treating or preventing KS. Even if the inhibition turns out to be far less effective in vivo compared to in vitro, such results may lead to discovery of some other proteins present in the cells (but absent in test tubes) or possibilities for other mechanisms involved in this receptor-ligand scenario.

Secondly, the researchers noted in their discussions that the lack of complete inhibition of infection by soluble a3b1 integrins suggest that a3b1 integrins are not the only receptor proteins. They may be able to utilize the gene that encodes for a3b1 integrins (I assume they already have this information, since plasmid constructs were used to express human a3 integrins in hamster cells), to search for homologues or similarities in other organisms, especially those known to be infected by HHV-8. Since we cannot use humans as our exprimental subjects to see if they develop KS or not, if the researchers find very similar mechanism and integrins at work in mice or hamsters, such discoveries may enable more in vivo experiments. Such in vivo experiments may help to understand how HHV-8 facilitate KS pathogenesis.

Lastly, the enzyme cascade involving focal adhesion kinase (FAK) can be further investigated. Although we have learned that blocking HHV-8 infection also blocks phosphorylation of FAK, we still don’t know what this phosphorylation exactly acheives in the cells. The antibodies used to detect phosphorylation in Figure 7 was anti-phospho-FAK antibodies, specifically designed to detect phosphorylation of only the FAK. Since enzyme cascades usually involve multiple phosphorylations in different proteins involved in the mechanism, I wonder if we can screen for any other proteins in the cells that are usually not phosphorylated, but specially phosphorylated due to the HHV-8 infection. I think that if such phosphorylated proteins are indeed found other than FAK, then may be the elusive enzyme cascade may be not too far from understanding.

1Purves, W.K., Orians, G.H., Heller, H.C., Sadava, D. (1998). Life. The Science of Biology. 285.
2Purves, 424.
3On-lineMedical Dictionary, Kaposi Sarcoma, 1997. http://cancerweb.ncl.ac.uk/cgi-bin/omd?query=Kaposi+Sarcoma