1. Introduction
A new promising approach to antitumor immunotherapy is the loading onto the tumor cell surface molecular structures recognized by innate immunity cytotoxic effectors such as natural killers (NK). There is ample evidence that 70 kDa heat shock proteins (HSP70) are such “elimination markers” capable of activating NK cells and other cytotoxic lymphocytes of the immune surveillance system [
1]. Tumor cells do not expose HSPs on their surface in a steady-state condition, but they express it under stress such as a heating.
Previously, we developed the two-component system to deliver HSP70 to the tumor cell surface consisting of the chimeric recombinant proteins 4D5scFv-dibarnase and barstar-HSP70 or barstar-HSP70_16, where HSP70_16 is a 16 kDa C-terminal fragment of HSP70 protein [
2]. The first targeting module was constructed from a 4D5 antibody fragment (scFv), binding the extracellular domain of HER2/neu receptor [
3], and fused with two molecules of a bacterial ribonuclease barnase [
4]. The second “effector” module contained barstar, a natural inhibitor of barnase. Barstar and barnase form heterodimers with
KD~10
−14 M [
5], exhibiting extraordinary stability in severe conditions. The specific binding of HSP70 to the HER2/neu-expressing tumor cells in vitro had a significant stimulating effect on the cytotoxic activity of NK cells [
2].
Recently, scaffold proteins such as monobody, anticalines, affibodies, or darpins have been considered worldwide as an alternative to targeting antibodies [
6,
7]. In particular, the HER2-specific scaffold protein of non-immunoglobulin nature DARPin9_29 (short for designed ankyrin repeat proteins) has been widely used in recent years as an HER2/neu-targeting molecule instead of the 4D5scFv [
8]. DARPin9_29 is an 18 kDa protein with a binding constant to HER2 of around 3.8 × 10
9 [
9]. Due to its relatively small molecular mass and, consequently, lower immunogenicity and higher ability to penetrate tissues, darpins are more attractive for therapeutic usage than antibodies. Darpin molecules are highly stable, and the absence of disulfide bonds allows their production in a bacterial expression system in high yield.
In this work, we developed an HSP70 delivery system using recombinant DARPin9_29-barnase protein as the antigen-recognizing targeting module and barstar-HSP70 or barstar-HSP70_16 proteins as the second one. For the comparison, we also produced and studied the hybrid single-module DARPin9_29-HSP70 protein.
2. Materials and Methods
2.1. Construction of DARPin9_29-HSP70 Expression Plasmid
To produce DARPin9_29-HSP70 vector, the DARPin9_29 sequence was amplified from pDARPin9_29-mCherry [
10] using 5′-tattccatatggacctgggtaagaaactg and 5′-ctgacagaattcggcgccttgcaggatttcagccag primers. Both DARPin9_29 amplicon and plasmid pET22-His6-barstar-HSP70 [
2] were digested with
NdeI and
SpeI, and the DARPin9_29 gene was cloned into the pET22 vector in the same reading frame with HSP70. The resulting expression cassette (
Figure 1a) consisted of an inducible T7 promotor, hexahistidine tag, and DARPin9_29 and HSP70 genes. The DARPin9_29 and HSP70 were connected by a flexible linker from the mouse IgG3 hinge region [
3]. After sequence verification, the plasmid pET22_His6-DARPin9_29-HSP70 was used for the protein expression. The recombinant proteins DARPin9_29-barnase, barstar-HSP70, and barstar-HSP70_16 were produced as described in [
2,
11].
2.2. Expression and Purification of DARPin9_29-HSP70
Freshly transformed E. coli BL21 (DE3) cells were grown in SOB medium containing 0.1 g/L ampicillin; the lac-promoter was induced with 1 mM ITPG at OD600 of 0.7. Expression was allowed to continue for 12 h at 37 °C. Cells were harvested by centrifugation at 6000× g for 15 min at 4 °C, and the cell pellet was resuspended in lysis buffer (20 mM Tris-base, 20 mM NaCl, 20 mM MgCl2, and 1 mg/mL lysozyme). The suspension was incubated for 30 min at room temperature and then sonicated on ice. Cellular debris was removed by centrifugation at 15,000× g for 30 min at 4 °C.
The solution was applied to a Ni2+-NTA column (GE Healthcare) equilibrated in 100 mM NaH2PO4, 10 mM Tris-HCl, 8 M urea, and pH 8.0 buffer (B1). DARPin9_29-HSP70 was eluted with pH 4.5 B1 buffer. The protein yield was 10 mg/L of growth medium.
2.3. PAAG Gel-Electrophoresis
Whole lysate of recombinant E. coli and purified DARPin9_29-HSP70 (30 µg/line) was denatured at 95 °C in Laemmli sample buffer for 5 min under reducing conditions (2-mercaptoethanol). Proteins were separated by 10% PAAG gels using a Bio-Rad (Singapore) electrophoresis system. Gels were stained with PageBlue™ Protein Staining Solution (Thermo Scientific, Waltham, MA, USA). Molecular weight (MW) standard PageRuler™ unstained Protein Ladder, 20 to 250 kDa (Thermo Scientific, Waltham, MA, USA), was used to verify the MW of the proteins.
2.4. Cells
Human HER2-positive adenocarcinoma cells SK-BR-3 and BT-474 were grown in RPMI-1640, supplemented with 10% fetal calf serum (FCS) and pen-strep-glut (all from PanEco, Moscow, Russia) (complete culture medium) in a CO2 incubator at 37 °C. Adhesive cells were passaged using Trypsin /EDTA solution (PanEco, Moscow, Russia) twice a week.
2.5. Interaction of the Dual-Module Molecular Constructs with HER2-Positive Cells
The SK-BR-3 and BT-474 cells were detached using Versen solution and transferred to a phosphate–salt buffer (PBS) containing 1% BSA and 0.1% NaN3, pH 7.4 (PBA). The 3 × 105 cells were then incubated with 100 μL of DARPin9_29-barnase (20 μg/mL, 40 min), washed twice with PBS and incubated with 100 μL of barstar-HSP70 or barstar-HSP70-16 (50 μg/mL, 40 min at +4 °C) and washed again. The single-module DARPin9_29-HSP70 was used at a concentration of 50 μg/mL. The negative control was treated in the same way but contained no constructs.
2.6. Flow Cytometry
Cell lines BT-474 and SK-BR-3 were pretreated with the constructs DARPin9_29-barnase/barstar-HSP70 or DARPin9_29-barnase/barstar-HSP70-16, washed and labeled with the antibody to HSP70, BRM-22 at 10 μg/mL, followed by the second antibody to murine IgG-FITC (both from Sigma, Merck KGaA, Darmstadt, Germany) at a dilution of 1:1000 in PBA for 40 min at +4 °C. The cells were analyzed by a FACSCalibur cytometer (BD, Franklin Lakes, NJ, USA) using propidium iodide to exclude dead cells. Data were analyzed by the FlowJo program, version 10 (BD, Franklin Lakes, NJ, USA).
2.7. Confocal Microscopy
SK-BR-3 and BT-474 cells were seeded on sterile cover glasses (15 × 104) placed into 6-well plates (Costar, Washington, DC, USA) and incubated overnight. Afterward, the cells were washed in PBA and incubated with DARPin9_29-barnase (20 μg/mL 40 min) and barstar-HSP70 or barstar-HSP70-16 (50 μg/mL, 40 min) and DARPin9_29-HSP70 (50 μg/mL, 40 min) followed by BRM-22 and anti-mouse IgG-Alexa Fluor 488 (Molecular Probes, Eugene, OR, USA) for 40 min at +4 °C. All the components were introduced sequentially with the wash in-between, stained with nuclear dye Hoechst 33,342 (Sigma, Merck KGaA, Darmstadt, Germany) and, finally, fixed with 2% paraformaldehyde and polymerized with Mowiol 4.88 medium (Calbiochem, Darmstadt, Germany). Slides were analyzed using an Eclipse TE2000 confocal microscope (Nikon, Tokyo, Japan). EZ-C1, version 3.90 FreeViewer software (Nikon Corporation, Tokyo, Japan) was used to process the results.
2.8. Isolation of Human Blood Mononuclear Cells
Peripheral blood mononuclear cells (PBMC) were obtained from the blood of healthy anonymous volunteers who gave their informed consent for participation in the study. All studies involving human cells were conducted in accordance with the guidelines of the World Medical Association’s Declaration of Helsinki. According to the national regulations, no additional approval by the local ethics committee was required in the case of anonymous blood cells being discarded after the experiment. PBMCs were isolated by density gradient centrifugation. The cells were then washed twice with Dulbecco PBS (DPBS) for 15 min at 300 g. The isolated PBMCs were then resuspended in magnetic separation buffer (DPBS, 2 mM EDTA, and 0.5% BSA).
2.9. NK Cell Isolation
NK cells were isolated by negative magnetic separation from the PBMCs using a MACS NK Cell Isolation Kit (Miltenyi Biotec, Bergisch Gladbach, Germany). After isolation, NK cells were transferred into complete culture medium and incubated overnight in a CO2 incubator at 37 °C. Human recombinant interleukin 2 (IL-2) (NPK BioTech, Sankt-Peterburg, Russia) was added to NK cells (500 U/mL) for 12 h before the experiments. A FACSVantage DiVa cell sorter (BD, Franklin Lakes, NJ, USA) was used to sort CD57+ and CD57− NK cell subsets. Sorted cells were transferred into culture medium and used in the experiments.
2.10. γδT Lymphocyte Isolation
Human γδT lymphocytes were isolated from PBMCs by positive magnetic separation using a gamma/delta T cell kit (MACS Mitenyi, Waltham, MA, USA). After separation, the cells were cultured for 6 to 7 days in complete culture medium supplemented with 10 ng/mL IL-2 and IL-4.
2.11. Lactate Dehydrogenase Cytotoxic Test
The CytoTox96 non-radioactive cytotoxicity assay (Promega, Fitchburg, MA, USA) was used to analyze the cytotoxic activity of NK cells. For this assay, SK-BR-3 cells were grown in a 96-well plate overnight (104/well). The cells were treated in PBS with the different modules as described above and finally transferred to complete culture medium. PBMCs or γδT cells were added to the treated cells at 10:1, 20:1 (PBMCs), and 5:1 (γδT cells) ratios. The plates were centrifugated to assist cell contacts and incubated in the CO2 incubator for 3 h. During the incubation, a lysing buffer was added to some wells 45 min before the end of the process as a control. At the end of the incubation, the plates were centrifuged, and the supernatant was transferred to another plate with the substrate mixture containing lactate, diaphorase, tetrazolium salts, and NAD+. The plates were incubated for 30 min. The concentration of lactate dehydrogenase was determined by converting tetrazolium salts into formazan. Cytotoxicity was analyzed as the ratio of the optical densities of the samples with a mixture of SK-BR-3 and NK cells to the sample with fully lysed targets, taking into account corrections for the presence of lactate dehydrogenase in the medium and the spontaneous release of lactate dehydrogenase from NK cells and SK-BR-3 cells. Optical density was measured using a Multiscan FC (Thermo Scientific) at 490 nm.
2.12. NK Cell Degranulation Cytotoxicity Test
Cytotoxicity was analyzed by assessing the expression level of past degranulation marker LAMP-1 (CD107a) on the surface of NK cells. NK cells were added at a 10:1 ratio to the SK-BR-3 cells treated before with the recombinant proteins as described above. Mouse anti-human monoclonal CD107a-APC (clone REA792, Miltenyi Biotec) antibody diluted in PBA (10 μg/mL) and brefeldin A (10 μg/mL) was added to stain NK cells and prevent LAMP-1 reuptake accordingly. The plate was centrifuged at 450× g for 2 min to assist cell contacts and incubated in 5% CO2 at 37 °C for 2.5 h. The plate was then centrifuged and washed twice with PBS. Analysis was performed using a FACSCalibur flow cytometer (Becton Dickinson, San Jose, CA, USA). Data processing and statistical analysis were performed in the FlowJo program, version 10 (BD, Franklin Lakes, NJ, USA).
2.13. Statistical Processing of the Obtained Data
In this study, the paired Student’s t-test was used for statistical processing of the results. The results of five independent experiments were evaluated with a statistical significance level of p < 0.05.
4. Discussion
HSP70 has long attracted the attention of researchers for use in antitumor therapy, but in the vast majority of cases, they are referring to complexes of HSP70 with tumor-specific peptides, which can be used to form tumor-specific immunity [
13,
14,
15]. We propose a fundamentally different approach that is aimed at the stimulation of the immune surveillance system, primarily at the development of a universal antitumor response of cytotoxic effector cells of innate immunity.
The antitumor response of cytotoxic effector cells of innate immunity (NK cells) associated with the localization of HSP70 on the surface of cancer cells was studied and described in detail by G. Multhoff [
16]. This team continues to develop this area of research at present, but in the applied aspect of antitumor immunotherapy, they proposed not the delivery of exogenous HSP70 to the surface of target cells but the use of monoclonal antibodies recognizing membrane-associated HSP70 [
17]. Thus, we believe that HSP70 is a promising target for the creation of molecular constructs, designed for modification of the surface of tumor cells to enhance antitumor cytotoxic immune response.
Targeted delivery of “cytolytic markers” can be accomplished by including in the recombinant constructs a guide module to other cancer markers, different from HSP, expressed on the surface of a wide range of malignancies. Experiments with targeted delivery of exogenous HSP70 to tumor-associated antigens are described by Poznanski’s group [
18]. In that study, a recombinant protein containing a mini-antibody to mesothelin, which is highly expressed in a mouse model of ovarian tumor, and mycobacterial HSP70 was produced and tested.
In this study, we use an original method of targeted delivery of HSP70, directed to the tumor antigen HER-2/neu, to the surface of tumor cells using a two-component construct capable of forming a supramolecular complex through HER2-DaRPin9_29 and barnase:barstar pair interactions. In this system, the function of the first module is to target binding to the surface of cancer cells. In our previous work [
2], we used as a recognizing HER2-receptor module the scFv 4D5-dibarnase, which consists of two barnase molecules that are fused serially to the single-chain variable fragment (scFv) of humanized 4D5 antibody (30 kDa). The present work focuses on DARpin9_29, which is an artificial scaffold protein designed as an alternative to antibodies. The advantages of this so-called alternative binding protein include its small size (18 kDa), which facilitates tumor penetration, the absence of Fc antibody- and complement-mediated cytotoxicity, and the high thermostability that enables its long-term storage at room temperature without loss of activity. The simplicity of production of DARpin9_29 fusion proteins makes them promising for creating bispecific and multivalent constructs. Our confocal microscopy (
Figure 2) and flow cytometry data (
Figure 3) showed that DARPin9_29 in the dual-module system is as effective as the antibodies. The delivery of both the HSP70 and HSP70_16 molecules with DARPin9_29 also significantly enhanced the antitumor cytolytic effect of PBMC (
Figure 4).
The barnase exposed on tumor cells serves as a site of selective binding of the second module consisting of barstar and HSP70 (or a fragment thereof). The selective interaction between the first and second modules is facilitated by the uniquely high binding constant of barstar to barnase; this protein heterodimer forms a complex with a K
D~10
−14 M, comparable only to the streptavidin-biotin system (K
D~10
−14 M). Our studies demonstrated the effectiveness of the barnase:barstar complex for the delivery of various drugs to cancer cells [
2,
3,
19].
The originality of the proposed immunotherapy method is also related to the application of a two-component molecular construct. Two-stage sequential treatment of tumor cells has an undoubted advantage in the delivery of HSP70 to tumor tissues in comparison with a single-module DARPin9_29-HSP70 (
Figure 2 and
Figure 3). It is known that HSP70 is able to form, using its substrate-binding site, strong complexes with the denatured proteins present, in particular, on the surface of stressed, damaged, dead, and transformed cells [
13]. Obviously, there is competition between the guiding part of the construct (mini-antibody, darpin, etc.) and the substrate-binding site in the HSP70 molecule for interaction with different target cells, and this will undoubtedly lead to the decreased efficiency of the targeted delivery of HSP70 to the cancer cell surfaces. The interaction of barstar with barnase is characterized by an association constant that is much higher than the constant for the interaction of HSP70 with the substrate molecules that may be present on the surface of other cells that are not targets of the current therapy. When using a two-component system, it is also possible to regulate the place, time, and number of administered drugs to maximize the therapeutic effect and reduce the number of adverse reactions.
In addition, according to our data, treatment of cells with the single-module DARPin9_29-HSP70 yields a much lower concentration of HSP70 on cell membranes compared to the dual-module system, which is likely a consequence of the lower affinity to the HER2/neu receptor of this chimeric protein due to steric hindrance created by the massive HSP70 molecule (
Figure 2 and
Figure 3). The same is true when comparing in our delivery system the second modules containing either the full-length HSP70 protein or its C-terminal fragment with a molecular mass of 16 kDa. The maximum fluorescence intensity determined by flow cytometry (
Figure 3) and proportional to the number of bound molecules was observed for the 16 kDa fragment of HSP70. However, the full-length protein HSP70-labeled cells are more efficiently lysed by PBMC and isolated γδT lymphocytes (
Figure 4 and
Figure 5).
We have demonstrated in our models the cytolytic response of cytotoxic immune effectors to membrane-associated HSP70 (
Figure 4 and
Figure 5). The results of the experiments using PBMC as effector cells demonstrated a reliable anti-cancer effect of the treatment of tumor cells with the two-module construct. This small effect was recorded in experiments with cytotoxic γδT lymphocytes, and a similar trend was observed in experiments with NK cells isolated from peripheral blood. The latter can be explained by the high level of HSP70-independent cytotoxic activity of NK cells against the target tumor cells. As for PBMC, we hypothesize that T lymphocytes may have a cytotoxic effect on tumor cells with membrane-associated HSP70. At the same time, we do not exclude the involvement of antigen-presenting cells contained in PBMCs in the development of the lytic reaction of T lymphocytes. Additionally, antigen-presenting cells can also enhance the cytotoxic response of NK cells by produced cytokines or/and by the interaction of classic and non-classic HLA class I molecules with the NK cell-activating receptors. The exact mechanisms of the cytotoxic action of NK cells and other cytotoxic lymphocytes against target cells carrying such constructs require further studies.
It is undoubtedly important to search for the region of the HSP70 molecule that induces the maximum level of cytolytic response of NK cells and other cytotoxic effectors of the immune system, since the response of cytotoxic cells used in this work differed in its specificity. It follows from published experimental data [
20] that the C-terminal domain of HSP70 is responsible for the stimulation of the cytotoxic and proliferative activities of NK cells. However, the results of our studies using a panel of monoclonal antibodies, which contains antibodies selectively interacting with the C- or N-terminal fragment of HSP70 [
21], indicate that membrane-associated HSP70 is predominantly recognized on many types of tumor cells by antibodies interacting not with the C- but with the N-terminal part of these proteins. This indicates a C-terminal interaction of the HSP70 with the cell surface, leading to an orientation to the outer space of the N-terminal portion of this molecule. Significantly, such target cells were efficiently recognized and subjected to cytolytic attack by cytotoxic effectors of the immune system [
22]. The influence of the spatial orientation of HSP70, as well as the search for regions of the molecule that induce cytotoxicity in different populations of cytotoxic lymphocytes, requires further investigation.
5. Conclusions
In the present study, we developed the two-component system for the delivery of HSP70 to the surface of HER2-positive tumor cells. This system is based on chimeric recombinant proteins, where the first antigen-recognizing module DARPin9_29-barnase contained a scaffold protein of DARPin9_29 as a guiding molecule to the HER2/neu receptor and a ribonuclease barnase. The second module contained a specific inhibitor of barnase barstar and HSP70 (barstar-HSP70) or its C-terminal 16 kDa region (barstar-HSP70_16) responsible for stimulation cytotoxicity of innate immunity cells. The specific binding of HSP70 with SK-BR-3 and BT474 tumor cells in vitro increased the cytolytic activity of PBMCs and γδT cells. For the comparison, we produced a single-module chimeric protein, DARPin9_29-HSP70, but it was much less efficient than the dual complex.
Thus, we assume that the targeted modification of the tumor cell surface with molecular structures that are recognized as “elimination markers” by cytotoxic effectors of the innate immune system is among the new promising approaches to antitumor immunotherapy. The exact mechanisms of cytolytic action of NK cells and other cytotoxic lymphocytes, as well as the search for regions of the HSP70 molecule responsible for this effect, require further studies.