In this report, we present evidence showing that the peptide S20-3, corresponding to the Ig-like domain of the Fas-targeting K1 protein of HHV-8, selectively kills hematological cancer cells, and the mechanism involves the Fas and TNFRI receptors. The cell-killing effect appears to be selective for cancer cells in vitro. In vivo, even a single intratumoral dose of peptide was active against the growth of xenograft tumors.
From the array of K1 Ig-like domain peptides tested (Table 1), only the S20-3 peptide demonstrated strong and reproducible cell-killing activity (Figure 1 and Figure 2) in all 6 hematological cell lines tested but not in PBMC controls (Figure 2). While it is not clear as to why S20-3, and also less reproducibly S20-2, but not other K1 Ig-like domain-derived peptides, possess cell-killing activity, the structural features of the predicted Ig-domain (Figure 5B) reveal a unique feature of the S20-3 peptide; a loop (centered at conserved glycine residue) linking 2 beta sheets, which are predicted to be destabilized or absent in the rest of peptides tested (Table 1). A truncated version of the S20-3 peptide, S10-1, representing the first beta sheet and the loop (Figure 5B), as well as S8-2 peptide, representing the second beta sheet (Figure 5B), lack cell killing properties (Figure 1B). On the other hand, a TCR-derived peptide sharing 5 structure-defining residues with S20-3 (Figure 5A) also showed cell-killing effect (Figure 5C), suggesting that the biological effect of S20-3 is related to its structure.
A seemingly contradictory effect of the whole Ig-like domain in K1 protein and S20-3 peptide on Fas signaling may also be explained by the structure-function relationship. The fact that peptide S10-1, but not S20-3 or any other K1 peptide, was able to disrupt the K1-Fas complex (Additional file 1: Figure S2) suggests that first beta sheet is involved in K1-Fas interaction. This is further supported by the fact that peptide S10-2, lacking 3 residues from the first beta sheet, failed to displace K1 (Additional file 1: Figure S2) and did not show any enhancement of FasL activity (Figure 1A). Additionally, peptide S20-2, which also contains S10-1 residues, showed cell-killing properties similar to peptide S20-3, but with reduced reproducibility, suggesting that the second beta sheet in peptide S20-3 increases structural stability of the peptide and the additional residues, preceding (S20-2) or following (S20-3) S10-1 region, affect peptide behavior. Taking all this into account, we hypothesize that the smaller size and possible flexibility of the loop within S10-1peptide as compared to S20-3 peptide (Figure 5B) allow access of this peptide to the K1 binding site and, thus, displacement of K1 from Fas (Additional file 1: Figure S2). The second beta sheet in the S20-3 peptide stabilizes the loop, but at the same time, it decreases loop flexibility and increases bulkiness of the peptide, limiting its access to the K1 binding site in the presence of the K1 protein.
This hypothesis also helps explain the differential effects of the K1 Ig-like domain, S10-1, and S20-3 on Fas receptor activation. The S10-1 sequence within the Ig-like domain in the whole K1 protein is flanked by additional domains of K1 protein. Assuming the S10-1 region within K1 is exposed and available to bind Fas, the limitations of the movement imposed by surrounding K1 domains “lock” the Fas receptor in the closed conformation, preventing binding of FasL described previously . On the other hand, the beta sheet and flexible loop in the S10-1 peptide can also bind the receptor, but without the rigidity of surrounding structures, its binding does not affect receptor conformation. Therefore, the S10-1 peptide has no direct effect on the receptor on its own, but sensitizes K1-positive cells to FasL (Figure 1A) by displacing the K1 protein (Additional file 1: Figure S2). The S20-3 peptide, more rigid and bulkier that S10-1peptide, can bind Fas only in the absence of K1. Without the flanking domains of the K1 protein and the whole Ig-like domain, S20-3 (and S20-2) can bind Fas receptor similarly to S10-1, but the presence of additional residues/structures induces conformational change mimicking the active state of the receptor.
The extrinsic apoptotic pathway involves activation of death receptors, recruitment of FADD, cleavage of pro-caspase-8, activation of caspases' cascade, and a drop in mitochondrial membrane potential . While the precise target for the cell-killing activity of the S20-3 peptide is unclear, data presented here clearly show that the peptide activates caspase-8, -9, and -3 (Figure 1D) and decreases mitochondrial membrane potential (Additional file 1: Figure S1), suggesting involvement of a death receptor, such as Fas. However, a conventional dose of the pancaspase inhibitor z-VAD blocked cell killing only incompletely (Figure 3B), and Jurkat cells with mutated inactive caspase-8 or dominant-negative FADD also showed only partial blockage of S20-3–induced cell-killing (Figure 3A), despite their inability to form the death-inducing signaling complex (DISC) . This persistence of the S20-3 peptide to kill mutant Jurkat cells (Figure 3A), the killing of Daudi cells that are considered Fas-resistant [17, 24], the increase of necrotic death in the z-VAD-treated Daudi cells (Figure 3C and Additional file 1: Figure S3A), and their relatively fast killing [necrotic cell death in Daudi cells was detectable 1 hour after peptide exposure (Additional file 1: Figure S3)] suggested to us that S20-3 also activates a TNF receptor.
Even though Fas belongs to the TNF receptor family and shares a significant structural similarity with TNFR , the outcomes of activating these receptors can be quite different . For example, activation of Fas receptor in L929 cells triggers apoptosis, whereas activation of TNFR triggers necrosis . Owing to the structural similarity, TNF-α is able to also bind and activate the Fas receptor . We, thus, investigated the possibility that, because of the structural promiscuity (further supported by the killing properties of a structurally related TCR peptide), the S20-3 peptide designed to bind the Fas receptor may also bind TNFR and trigger necrosis. We detected TNFRI expression in BJAB, Jurkat, and Daudi cells (Figure 3), and the TNFRI-blocking antibody significantly inhibited S20-3– and TNF-α–induced cell killing in all 3 cell lines (Figure 4B and C). On the contrary, the TNFRII-blocking antibody showed no inhibitory effect on the S20-3 cell-killing of TNFRII-positive Daudi cells (Figure 4B). This finding is not surprising considering the fact that activation of TNFRII triggers pro-survival signaling in hematological cancer cells , and activation of TNFRI is required for any death signaling from TNFRII due to the lack of a death domain in TNFRII .
Our results with FADD– and caspase-8–defective Jurkat cells are in agreement with the reports showing that under apoptosis-deficient conditions (such as non-functional caspase-8 or FADD), stimulation with FasL or TNF-α could induce cell death with morphological features of necrosis/necroptosis [21, 28, 29]. Furthermore, lack of FADD, but not of caspase-8, was shown to sensitize Jurkat cells to TNF-α–induced necrosis . Smac mimetic BV6 enhanced TNF-induced cell death in leukemia cells in 2 different ways: necroptosis, when the cells were apoptosis resistant (FADD– and caspase-8–deficient), and caspase-8–dependent apoptosis in apoptosis-proficient cells .
We hypothesize that the different death pathways can be activated in response to S20-3 treatment in Jurkat, Daudi, and BJAB cells, depending on the availability of and sensitivity to Fas and TNFRs. Another possibility is a cross talk between signaling events from TNF and Fas receptors, as reported by Takada et al., in which TNFRI is recruited by Fas to induce apoptosis .
An additional important observation is that the S20-3 peptide activity seemed to be specific to malignant cells; leukemia T cells displayed a much greater sensitivity to S20-3 than nonmalignant cells (Figure 2C). While the constitutive expression of TNF receptors was clearly demonstrated in most tumor cells, in normal peripheral lymphocytes, the expression of TNF receptors is subjected to a positive and negative regulation and can be induced by different stimuli [33, 34]. However, normal unstimulated PBMCs express very low amounts of mRNAs for TNFRII > TNFRI > Fas , and normal lymphocytes were shown to be resistant to stimulation with activating antibodies targeting TNFRI, TNFRII, or Fas . Thus, our findings of cancer-specific killing by the S20-3 peptide are in agreement with these reports.