Figure 4
The substitution of Ser52 by an Arg residue has a tremendous impact on
the binding affinity of B7-H6. In this mutation, a hydrophobic amino
acid residue is replaced by a positively charged one, which completely
changes the polarity of the protein surface, dramatically increasing the
dissociation constant (KD). Replacement of Ser52 with
another hydrophobic amino acid (Ala in this case) has a much more
reduced impact on the binding of B7-H6. Replacement of Phe 85 with a
non-aromatic amino acid residue causes an accentuated decrease in the
pocket hydrophobicity. The B7-H6 region that interacts with this portion
of NKp30 comprises the residues Thr59 and Pro128, each making sixteen
van der Waals contacts with NKp30, that account for 44% of all contacts
between the two proteins (Li, Wang & Mariuzza, 2011). This region of
B7-H6 forms a projection that seems to bind at the groove formed between
Ser52 and Ser82 of NKp30 (see Figure 4) , which may be a key
feature involved in triggering signal transduction by NKp30.
Further work yielded the X-ray structure of NKp30 bound to B7-H6,
providing important information regarding the interaction of these two
molecules. Interestingly, B7-H6 interacts with just eleven NKp30
residues, mainly through hydrophobic interactions. The three potentiallyN -glycosylated residues identified by other authors (Asn residues
42, 68 and 121) are located outside the interface and therefore should
not interfere with binding (Figure 4 ) (Li, Wang & Mariuzza,
2011). However, it was argued that different glycosylation patterns may
affect the overall 3D structure of the protein, increasing or decreasing
binding affinity, as already discussed (Hartmann et al., 2012).
Marked NKp30 conformational changes occur upon B7-H6 binding, as
observed by comparison of the two available structures (bound and
unbound). A key change is the reduction of the distance between residues
52 and 82, that define the ridge where B7-H6 docks. Moreover, the loop
composed by Arg67 and Asn68, located above the ridge, seems to rotate,
pushing Asn68 backwards and pulling the Arg67 residue down towards the
centre of the ridge. Although the available structures do not allow the
identification of an H-bond, the NH moiety of the side chain of Arg67
approaches the C=O of Val53 at about 4Å.
Interestingly, Asn68 is one of the 3 identified glycosylation acceptor
sites. Although all these sites are located outside the binding site,
the glycosylation of Asn42 and Asn68 is essential for B7-H6 binding;
their glycosylation is likely to induce subtle NKp30 conformational
changes, helping shape the ligand-binding pocket (Hartmann et al.,
2012). On the other hand, BAT-3 binding depends essentially on Asn68
glycoslyation, whereas Asn42 and Asn121 have a lower impact(Hartmann et
al., 2012). This indicates that Asn68 is critical for correct ligand
binding and subsequent signal transduction.
Although no experimental information exists regarding these
conformational changes, it is clear that some degree of modification is
required for NKp30-based triggering of NK cells. One particular work
demonstrated that, upon binding of B7-H6, conformational changes in
NKp30 enable the translocation of Arg143 to a position deeper in the
phospholipid bilayer. With this, the positively charged arginine residue
aligns with the negatively charged aspartate in CD3ζ or FCεRIγ adaptor
proteins, that trigger the NK cell response (Memmer et al., 2016).
Targeting NKp30
B7-H6 is expressed by approximately 20% of available human tumour cell
lines and by some primary human tumours (Brandt et al., 2009). It has
been widely demonstrated that B7-H6 expression is directly correlated
with susceptibility to NK-killing, as engaging of NKp30 by this protein
triggers a strong cytolytic response. It has also been suggested that
some cancer therapies may increase the stress levels on tumour cells and
upregulate the expression of B7-H6, increasing the susceptibility to NK
cells (Cao, Wang, Zheng, Wei, Tian & Sun, 2015). Moreover, it has been
demonstrated that artificially coating cells with B7-H6 promoted NK cell
cytotoxicity, providing a proof of concept that NKp30 engagement may
represent a new strategy in cancer therapy (Kellner et al., 2012).
In 2013, a patent was filed describing a method of endowing T cells with
receptors that recognise B7-H6-expressing cells, as a novel cancer
immunotherapy (Zhang & Sentman, 2013). In that work a bi-specific T
cell engager (BiTE) antibody was produced by fusing an anti-B7-H6
antibody fragment with an anti-murine CD3 scFv (single chain variable
fragment of antibodies). BiTE was successful in triggering the response
of host T cells against B7-H6-expressing cells (Wu, Zhang, Gacerez,
Coupet, DeMars & Sentman, 2015). CAR-T cells bearing B7-H6-recognising
domains have also been described, demonstrating specific activity
against B7H6-expressing tumour cells (Hua, Gacerez, Sentman & Ackerman,
2017). These approaches, although efficient, require the expression of
B7-H6 on target cells, something that is tumour type-dependent, as
already referred. Strategies aiming at increasing the expression of this
protein in tumour cells have not been explored, probably because a
number of reports suggest a positive correlation of the protein levels
with tumour metastasis and progression (Xu et al., 2016; Zhang et al.,
2018; Zhou, Xu, Chen, Xu, Wu & Jiang, 2015). Therefore, direct
targeting of NKp30 has been considered in the development of NK
cell-based therapies.
One work focused on triggering the response of NK cells with
B7-H6-derived small peptides. In this approach, the peptide sequence
present in the projection of B7-H6 that binds to the ridge of NKp30 was
reproduced as a soluble fragment. Treatment of NK cells with this
molecule induced the release of TNF-α, proving that NKp30 may be
targeted using small peptides. This work, however, failed in achieving a
complete response of NK cells, as no release of IFN-γ was triggered by
the designed peptides and no data regarding the cytolytic activity of NK
cells was reported (Phillips, Romeo, Bitsaktsis & Sabatino, 2016).
Despite the advances in NK cell characterization and in the
identification of the triggering mechanisms, few therapies using NK
cells are available for cancer treatment. In some ongoing clinical
trials, autologous NK cells are expanded ex vivo and reinfused
after stimulation with cytokines, namely IL-2. In another approach,
allogenic NK cells are directly administered to the patients, and an
immortalized NK cell line (NK-92) is used in some instances. However,
most of the ongoing trials rely on the use of specific antibodies aimed
at triggering the cytolytic responses of NK cells against cancer cells,
by targeting NK cell receptors and cancer cell markers. Some of the
therapies were found to be effective and to significantly increase the
life quality of the patients, while others presented serious side
effects (Dianat-Moghadam, Rokni, Marofi, Panahi & Yousefi, 2018).
Only a few works have focused on NKp30 as a target for the activation of
NK cells and as a possible route towards the development of new
immuno-oncology therapies. With all the benefits of immunotherapies and
as part of the personalized medicine approaches, it is expected that, in
the future, new drugs and treatments focusing on harvesting the power of
the immune system through NK cells will be developed.