Novel Human Anti-PD-L1 mAbs Inhibit Immune-Independent Tumor Cell Growth and PD-L1 Associated Intracellular Signalling

The novel antibody-based immunotherapy in oncology exploits the activation of immune system mediated by immunomodulatory antibodies specific for immune checkpoints. Among them, the programmed death ligand-1 (PD-L1) is of particular interest as it is expressed not only on T-cells, but also on other immune cells and on a large variety of cancer cells, such as breast cancer cells, considering its high expression in both ErbB2-positive and Triple Negative Breast Cancers. We demonstrate here that PD-L1_1, a novel anti-PD-L1 T -cell stimulating antibody, inhibits PD-L1-tumor cell growth also by affecting the intracellular MAPK pathway and by activating caspase 3. Similar in vitro results were obtained for the first time here also with the clinically validated anti-PD-L1 mAb Atezolizumab and in vivo with another validated anti-mouse anti-PD-L1 mAb. Moreover, we found that two high affinity variants of PD-L1_1 inhibited tumor cell viability more efficiently than the parental PD-L1_1 by affecting the same MAPK pathways with a more potent effect. Altogether, these results shed light on the role of PD-L1 in cancer cells and suggest that PD-L1_1 and its high affinity variants could become powerful antitumor weapons to be used alone or in combination with other drugs such as the anti-ErbB2 cAb already successfully tested in in vitro combinatorial treatments.

www.nature.com/scientificreports www.nature.com/scientificreports/ Biological activity of PD-L1_1 on breast cancer cells. Since PD-L1_1 is able to recognize with high affinity and specificity PD-L1 expressed not only on T-cells but also on breast cancer cells, we decided to investigate the in vitro effects of PD-L1_1 on breast tumor cells. To this aim, PD-L1_1 was tested at increasing concentrations (50-200 nM) on mammary SK-BR-3 and MDA-MB231 cells for 72 hours at 37 °C in the absence of lymphocytes. As a control, PD-L1_1 was also tested in parallel, in the same conditions, on PD-L1-negative MCF-7 breast cancer cells. As shown in Fig. 1e, PD-L1_1 significantly inhibited the growth of both the PD-L1-positive cell lines in a dose dependent-manner, whereas no effects were observed on the viability of MCF-7 cells, thus confirming the specificity of its biological effects.
Furthermore, the antitumor activity of PD-L1_1 was also tested in comparison with that of an anti-mouse PD-L1 (clone 10F.9G2, BioXcell) on mouse CT26 colon cancer cells. They were both found able to inhibit cell viability of about 30-40% at a concentration of 200 nM (see Fig. 2), thus indicating that the antitumor effect of PD-L1-1 was exerted not only on mammary cancer cells but also on different types of tumor cells.
In order to compare the biological anti-tumor activity of PD-L1_1 with that of the clinically validated anti-PD-L1 mAb Atezolizumab, we tested them in parallel at the dose of 100 nM on the indicated breast cancer cells ( Fig. 3 and Supplementary Fig. S1), by including an unrelated IgG4 isotype antibody as a negative control. As a further positive control, two variants of PD-L1_1 with higher affinity for PD-L1, called 10_3 and 10_12 (Cembrola et al., Rapid affinity maturation of novel anti-PD-L1 antibodies by a fast drop of the antigen concentration and FACS selection of yeast libraries, submitted for publication, 2019) were tested in parallel assays. These affinity-matured anti-PD-L1 antibodies were obtained by yeast surface display FACS-based methodology coupled with a CDR-targeted mutagenesis protocol applied to a single CDR in the heavy chain of PD-L1_1. The best variant 10_3 and 10_12 IgGs show lower equilibrium dissociation constants (about 10 fold lower K D ) with PD-L1 than the wild type one (see Fig. 4). Accordingly, they were found able to strongly inhibit the growth of both the tumor cell lines, with effects even more potent than those of the parental PD-L1_1 and Atezolizumab. When tested on MCF-7 cells, used as a negative control, PD-L1_1 and its high affinity variants did not show significant effects ( Fig. 3 and Supplementary Fig. S1), as expected.
Unfortunately, 10_3 and 10_12 did not show the anti-tumor effects of the parental PD-L1_1 when tested on CT26 cancer cells due to unexpected loss of their binding ability to these cells (see Fig. 4 Effects of anti-PD-L1 mAbs on intracellular pathways downstream PD-L1. Considering the in vitro anti-tumor effects of the novel isolated anti-PD-L1 mAb and its high affinity variants on breast cancer cells we made the hypothesis that PD-L1 may play by itself a role on tumor cells, by inducing cell proliferation, and anti-PD-L1 mAbs might inhibit its effects. Effects of PD-L1_1 on MDAMB231 (gray curve), SK-BR-3 (black curve) or MCF-7 cells(dashed line). Cells were treated for 72 h with PD-L1_1 mAb tested at increasing concentrations and cell survival was expressed as percentage of viable cells with respect to untreated cells (e). Error bars depicted means ± SD. P values for the indicated treatment relative to untreated cells, are: ***P ≤ 0.001; **P < 0.01; *P < 0.05. www.nature.com/scientificreports www.nature.com/scientificreports/ To test this hypothesis on the role of PD-L1, we firstly used PD-1/Fc fusion protein as an agonist, to activate PD-L1 and eventually tumor cell growth, and IFN-γ to inhibit cell growth and induce apoptosis 10,30 .
After the treatments of SK-BR-3 cells with PD-1/Fc (1 µg/ml) or IFN-γ (100 ng/ml), carried out for 72 hours at 37 °C, we analysed the effects on both tumor cell survival and on the pathways downstream PD-L1.
As shown in Supplementary Fig. S2b, we observed an increase of cell proliferation when the tumor cells were treated in the presence of PD-1/Fc and a significant reduction of the cell viability when they were treated with IFN-γ, accordingly with similar effects of IFN-γ on other tumor cell lines previously reported 28,[30][31][32][33][34][35] . As a negative control, these experiments were repeated on PD-L1-negative MCF-7 cells and no significant effects were observed ( Supplementary Fig. S2a).
In parallel, by western blotting analyses we evaluated the level of phosphorylated Erk, which was reported to play a role as a promoter of tumor cell proliferation, and Cleaved caspase-3, as a marker of apoptosis 14,15 . As shown in Supplementary Fig. S2c, the level of p-Erk increased in the presence of PD-1/Fc, whereas the level of Cleaved caspase-3 significantly decreased compared to untreated cells. On the contrary, when the cells were treated with IFN-γ, the level of p-Erk was reduced, whereas the level of Cleaved caspase-3 was higher compared to untreated cells.
These results suggest that PD-L1, upon binding to its receptor PD-1, induces the phosphorylation of Erk and inhibits tumor cell death, in line with similar results observed in a PD-L1+ cell line, and previously reported 11,12 . Indeed, PD-L1 has been reported in literature to affect tumor cell proliferation by increasing the levels of p-Erk, p-JNK and p-P38 proteins, which are members of the MAPK family and play an important role in the transduction of extracellular signalling, thus regulating several cellular functions such as cell proliferation, survival and differentiation [13][14][15] .
In order to test whether the PD-L1_1 antibody inhibits tumor cell proliferation by similarly affecting these pathways, we performed western blotting analyses of extracts from breast SK-BR-3 and triple negative MDA-MB-231 tumor cells treated for 72 hours at 37 °C in the absence or in the presence of PD-L1_1, used at the concentration of 200 nM. As shown in Fig. 5a and in Fig. 5b, the level of phosphorylated Erk, P38 and JNK proteins significantly decreased when both SK-BR-3 and MDA-MB-231 cells were treated with PD-L1_1 compared to untreated cells. No significant effects were observed on the level of total Erk, P38 and JNK (data not shown), as well as no significant effects were observed on both the levels of p-Akt and total Akt (see Supplementary Fig. S3).
Atezolizumab, used as a positive control on SK-BR-3 tumor cells, showed similar effects on p-Erk, but, differently from PD-L1_1, did not affect the level of p-JNK, and showed only a slight effect on p-P38 (Fig. 5a). As additional controls, the two high affinity variants of PD-L1_1, 10_3 and 10_12, were tested in parallel assays on SK-BR-3 or MDA-MB-231 tumor cells. After treatments with 10_3 and 10_12, the phosphorylation levels of Erk, P38 and JNK were found to be strongly reduced by these variants (Fig. 5a,b), that showed a more potent effect compared to the parental PD-L1_1 mAb in both tumor cell lines. However, the effects obtained with the two variants of PD-L1_1 were more marked on SK-BR-3 cells compared to those observed on MDA-MB-231 cells, in line with the results described above on tumor cell viability, in which 10_3 and 10_12 inhibited tumor cell growth of SK-BR-3 tumor cells more efficiently compared to MDA-MB-231 tumor cells.
Thus, the anti-PD-L1 antibodies seem to stress tumor cells and inhibit their cell proliferation supporting the idea that PD-L1 could represent a marker for cancer, independent from the immune system. Furthermore we   To confirm this hypothesis on the role of PD-L1 on tumor cells, we tested also the effects of PD-L1_1 and a commercially available anti-mouse PD-L1 mAb (clone 10F.9G2, BioXcell), previously validated in vivo 36 , on these www.nature.com/scientificreports www.nature.com/scientificreports/ pathways in PD-L1-positive colon CT26 tumors in vivo. To this aim, mice were implanted with CT26 cells (day 0) and then treated with PD-L1_1 or anti-mouse PD-L1 antibody (200 µg ip, clone 10F.9G2, BioXcell) reacting against murine PD-L1 (day 3, 6, 10). While the growth rate of tumors in untreated mice was very fast and uncontrolled, with the majority of tumors reaching sizes of >650 mm 3 at day 21, a drastic reduction in tumor volume was found in mice treated with α-mPD-L1 (p = 0.02), and similar effects were observed in mice treated with PD-L1_1, as expected 9 (Fig. 6a). During the period of treatment, the animals did not show significant changes of weight or other visible signs of toxicity.
We then analyzed the effects of PD-L1_1 and α-mPD-L1 treatments on the activation of MAPK proteins having a critical role in cancer cell proliferation and found to be involved in PD-L1 intracellular signaling. Western blotting analyses were performed on cell extracts from tumors removed at the end of the experiment on day 21, and processed as described in Methods. As shown in Fig. 6b and in Fig. 6c, the anti-PD-L1 antibodies inhibited the phosphorylation/activation of MAPK and JNK and induced the cleavage of caspase-3 more efficiently compared to untreated tumor cells, thus confirming the results relative to the association of PD-L1 to these pathways observed in vitro, and mentioned above. Protein levels are also expressed as fold increase with respect to those observed in untreated mice and normalized to actin. Error bars depicted means ± SD. P values for the indicated mAbs relative to cell extracts from untreated groups, are: ***P ≤ 0.001; **P < 0.01; *P < 0.05. www.nature.com/scientificreports www.nature.com/scientificreports/ Antitumor effects of PD-L1_1 combined with anti-ErbB2 antibody. It has been reported that combinatorial treatments of anti-PD-L1 with anti-ErbB2 drugs can lead to more potent effects as compared to treatment with each single agent 8 (ClinicalTrials.gov Identifier:NCT03125928).
Due to the in vitro anti-tumor efficacy of the novel isolated PD-L1_1 mAb and to its ability to recognize with high affinity and specificity PD-L1 also on breast cancer cells, inhibiting cell growth and affecting the downstream MAPK pathways, we investigated the possibility to use it in combination with the anti-ErbB2 compact antibody, Erb-hcAb, capable of inhibiting breast tumor cell growth in vitro and in vivo 21,24,37 .
With this aim SK-BR-3 cells were treated for 72 hours at 37 °C with increasing concentrations (50-100 nM) of Erb-hcAb and PD-L1_1 mAbs, used alone or in combination. We show here that the combinatorial treatment inhibits the growth of tumor cells more efficiently than when they are used as single agents (Fig. 7). An unrelated IgG4 isotype control antibody was used as a negative control (data not shown).
To test whether the combinatorial treatment could be useful also for inhibiting the growth of more aggressive breast cancer cells resistant to Trastuzumab, we used JIMT-1 cells, derived from a metastasis of breast cancer patient, in a parallel assay. As shown in Fig. 8, the combinatorial treatment inhibited the growth of these tumor cells more efficiently than when they were used as single agents, proving to be beneficial also in Trastuzumab-resistant JIMT-1 cells.
The efficacy of the combinatorial approach was then further tested on co-cultures of SK-BR-3 or JMTI-1 tumor cells with hPBMCs to exploit also the effector functions of Erb-hcAb (ADCC mediated by Fc) and the inhibitory effects of PD-L1_1 mAb in the interaction of PD-1/PD-L1 21,38 . To this aim, we treated SK-BR-3 or JMTI-1 tumor cells with each antibody or with their combination at the concentrations of 25-100 nM in the absence or in the presence of hPBMCs (effector: target ratio 10:1) for 24 hours at 37 °C. As shown in Fig. 9, the presence of lymphocytes, as expected, potentiated the antitumor effects of both mAbs on SK-BR-3 cells.
However, the strongest effects were again observed when the two antibodies were used in combination, leading to total inhibition of cancer cell survival when the mAbs were tested at the highest dose of 100 nM (Fig. 9). An unrelated IgG4 isotype antibody was used as a negative control (data not shown).
The cytotoxic effects on target cells were also analyzed in a parallel assay, in which we evaluated the levels of LDH release from treated cells, as a measure of cell lysis 39 . Figure 9b shows that the highest level of LDH release was observed in the supernatant of cells treated with the combination of the two mAbs in the presence of lymphocytes.
Similar potentiated effects were observed when PD-L1-1 was tested in the presence of lymphocytes on JIMT-1 cells, showing additive effects on cell survival and lysis (LDH release) when it was used in combination with Erb-hcAb (data not shown). www.nature.com/scientificreports www.nature.com/scientificreports/

Discussion
Cancer immunotherapy, mediated by both antibodies and immune effector cells, is becoming a precious strategy to overcome the limits of the conventional therapies.
The comprehension of the mechanisms underlying T-cell proliferation and regulation improved the efficacy of the therapeutic approaches with the identification of antibodies agonistic for co-stimulatory or antagonistic for inhibitory receptors to enhance the potential of the immune-based cancer therapy.
Here we investigated the anti-tumor effects on breast cancer cells of a novel anti-PD-L1 antibody, PD-L1_1, generated from a large repertoire of fully human antibodies specific for many T-cell immune checkpoints (ICs) obtained by phage display technology, and capable of inducing T cell activation, cytokine secretion and antitumor effects in vitro and in vivo 9 .
Indeed, PD-L1 has been recently reported as a potential target for breast cancer, due to its high expression in both ErbB2-positive and Triple Negative Breast Cancer. A high proportion of PD-L1-positive tumors are infiltrated with PD-1-positive lymphocytes and, more interestingly, PD-L1 is expressed not only on T cells but also on breast cancer cells, thus it can be considered as a potential co-target for breast cancer treatments 26,27 .
We found that the novel PD-L1_1 mAb inhibited the growth of breast tumor cells expressing PD-L1 even in the absence of T-cells, thus confirming that PD-L1 plays also an additional role in tumor cells in promoting cell proliferation. We investigated also the effects of two high affinity variants obtained from PD-L1_1 (Cembrola et al., Rapid affinity maturation of novel anti-PD-L1 antibodies by a fast drop of the antigen concentration and FACS selection of yeast libraries, submitted for publication, 2019), and we found that they inhibited tumor cell viability more efficiently than the parental PD-L1_1. These results can be explained by considering that PD-L1 has been indeed reported in literature to induce tumor cell proliferation by increasing the levels of Ki-67, p-ERK, p-JNK and p-P38 11,12 , even though there are not yet clear evidences on a complete specific pathway downstream PD-L1.
Thus, we further investigated the anti-tumor potential of PD-L1_1 by evaluating its effects on intracellular pathways downstream PD-L1 in tumor cells. We demonstrated for the first time that the novel anti-PD-L1 mAb (PD-L1_1) can affect the phosphorylation of Erk, JNK and P38 in tumor cells. The high affinity variants of PD-L1_1 showed even more potent effects on the intracellular MAPKs, accordingly with their higher affinity for PD-L1 and their higher in vitro antitumor efficacy, thus confirming that these intracellular proteins are indeed regulated by PD-L1 on tumor cells. Further evidences supporting the association of PD-L1 with the mentioned pathways in tumors has been provided by the in vitro effects of Atezolizumab on p-Erk and p-JNK. The same activity was confirmed in vivo by using PD-L1_1 and a commercially available antibody recognizing the murine www.nature.com/scientificreports www.nature.com/scientificreports/ PD-L1. Unfortunately, this in vivo study was not carried out also with PD-L1_1 high affinity variants as they lost cross-reactivity with mouse PD-L1 after the affinity maturation process (see Fig. 4).
Currently, ongoing clinical trials in cancer patients include immunotherapy based on immunomodulatory antibodies used in combination with chemotherapy or anti-TAA mAbs 8,25 . On the basis of these considerations and of the results relative to the anti-tumor efficacy of the novel isolated PD-L1_1 mAb, we investigated the inhibitory effects of PD-L1_1 on tumor cells in combinatorial treatments with the anti-ErbB2 antibody, Erb-hcAb, capable of inhibiting tumor cell growth in vitro and in vivo 21 .
Thus, the anti-ErbB2 antibody, Erb-hcAb, was combined with PD-L1_1 mAb for the treatment of ErbB2-positive breast cancer cells, showing that the combinatorial treatment inhibits the growth of tumor cells more efficiently than when they are used as single agents on different tumor cell lines, including the aggressive Trastuzumab-resistant JIMT-1 cells. Even stronger anti-tumor effects were obtained when PD-L1_1 was used in combination with Erb-hcAb on breast cancer cells in the presence of lymphocytes. Indeed, we observed a total inhibition of cancer cell survival when the combination of the mAbs was tested at the dose of 100 nM.
Altogether, these results shed light on the role of PD-L1 in cancer cells and suggest that PD-L1_1 and its high affinity variants can become powerful antitumor weapons that can be used alone or in combination with other drugs, to achieve more potent antitumor effects.

Enzyme-Linked Immunosorbent Assays (ELISA).
To compare the binding affinities of PD-L1_1 and its variants,, ELISA assays were performed on chimeric PD-L1/Fc protein coated at 5 μg/mL on NuncTM flat-bottom 96-well plates (Thermo Fisher Scientific, 3596). in a solution of 0.05 M NaHCO 3 for 72 hours at 37 °C. After blocking of the coated 96-well plates with 5% nonfat dry milk in PBS for 1 hour at 37 °C, the purified mAbs were added at increasing concentrations to the plates in 2.5% nonfat dry milk in PBS and incubated for 2 hours at room temperature by gently shaking. After extensive washes with PBS, the plates were incubated with HRP-conjugated anti-human IgG (Fab')2 goat monoclonal antibody (Abcam, ab98535) for 1 hour, washed again and incubated with TMB reagent for 10 min before quenching with an equal volume of 1 N HCl.
To confirm the binding specificity of PD-L1_1, cell ELISA assays were performed on PD-L1-positive or PD-L1-negative breast cancer cells. Cell ELISA assays were performed by plating the cells in round-bottom 96-well plates (2•10 5 cells for each well) and incubating them with increasing concentrations (0.5-200 nM) of mAb in 2.5% nonfat dry milk for 2 hours at room temperature with gentle agitation. After the incubation with the primary antibodies, extensive washes were carried out with PBS, then the plates were incubated with an appropriate HRP-conjugated antibody for 1 hour at room temperature, washed again and incubated with 3,3′,5,5′-tetramethylbenzidine (Sigma-Aldrich, St. Louise, USA) reagent for 10 minutes before quenching with an equal volume of 1 N HCl. Absorbance at 450 nm was measured by the Envision plate reader (Perkin Elmer, 2102, San Diego, USA).
For After incubation on ice for 20 minutes, the extracts were clarified by centrifugation at 12000 rpm for 15 minutes at 4 °C. Protein concentration was determined by the Bradford colorimetric assay (Sigma-Aldrich, USA) and Western Blotting analyses were performed by incubating the membranes with anti-p-Erk, anti-p-P38, anti-p-JNK, anti-pAkt, anti-Akt or anti-Cleaved Caspase-3 antibodies, followed by the HRP-conjugated secondary antibody.

Isolation of human peripheral blood mononuclear cells (hPBMCs).
Human PBMCs were isolated, as previously reported 9 , by using Greiner Leucosep tube (Sigma-Aldrich, St. Loiuse, USA) following the manufacturer's instructions, and frozen in a solution containing 90% FBS and 10% dimethyl sulfoxide (DMSO) until use. Cryopreserved cell vials were gently thawed out by using RPMI 1640 medium supplemented with 1% L-glutamine, 1% CTL-Wash (Cellular Technology Limited, Shaker Heights, USA), and 100 U/mL Benzonase (Merck Millipore, Darmstadt, Germany). The collected hPBMCs were then washed by centrifugation, plated and incubated overnight at 37 °C in R10 medium consisting of RPMI 1640 supplemented with 10% inactivated FBS, 1% L-glutamine, 50 U mL −1 penicillin, 50 μg mL −1 streptomycin and 1% HEPES (GibcoTM, Thermo Fisher Scientific, Paisley, UK). After an overnight resting, the hPBMCs were collected in phosphate-buffered saline (PBS, Verviers, Belgium), counted by using the Muse ® Cell Analyzer (Merck Millipore, Darmstadt, Germany) and resuspended at a density of 1•10 6 cells/mL. www.nature.com/scientificreports www.nature.com/scientificreports/ Viable cells were counted by the trypan blue exclusion test and cell survival was expressed as percent of viable cells in the presence of the drugs under test with respect to negative control cultures grown in the absence of the proteins.

Cell viability and Cytolysis assays.
To test the effects of combinatorial treatments on co-cultures of tumor cells and hPBMCs, the cells were plated in 96-well flat-bottom plates at the density of 1.5•10 4 cells/well for 16 hours, hPBMCs from healthy donors were added at effector:target ratio 10:1 in the absence or presence of increasing concentrations of Erb-hcAb or PD-L1_1 mAbs, used alone or in combination (25-100 nM), and incubated for 24 hours at 37 °C. Controls included target cells incubated in the absence of effector cells or in the presence of the immunoagents alone.
After the treatment, lymphocytes were removed and adherent cells were washed and counted by the trypan blue exclusion test. Cell survival was expressed as percent of viable cells in the presence of the proteins under test with respect to the untreated cells, used as negative control.
Tumor cell lysis was determined by measuring the release of lactate dehydrogenase (LDH) in the supernatant of co-cultures described above by LDH detection kit (Thermofisher Scientific, Rockford, USA), following the manufacturer's recommendations. Lysis was calculated by measuring the fold increase of LDH in the presence of each mAb, with respect to the amount present in the supernatant of untreated cells, used as a negative control.
Typically, cell survival and cytolysis values were obtained from at least three independent experiments in which triplicate counts were determined. The images of the cells untreated or treated with each compound or with their combinations were acquired by a Leica Microsystems integrated microscope (DFC320, Cambridge, UK).
In vivo studies on mouse models. Six-weeks old female BalBC mice (Envigo, USA) were used for in vivo studies. Mice were challenged with a subcutaneous injection of 2•10 5 CT26 cells (day 0). Three days after, mice were left untreated (control) or treated with PD-L1_1 (200 µg ip) or anti-mouse PD-L1 (200 µg ip, clone 10F.9G2, BioXcell) administered at day 3, 6 and 10. Tumor growth for individual mice was monitored over time using a digital caliper every 3-4 days up to day 21. Tumor volume was calculated by using the formula: 0.5 × length × width 2 .
Tumors from control group and mice treated with PD-L1_1 were harvested at day 21, subjected to three homogenization cycles (3 minutes at 30 Hz) by using RIPA buffer, containing Protease inhibitors and Na 3 VO 4 , and centrifuged. For the analysis of tumor cells in the absence of infiltrating lymphocytes, tumors from control group and mice treated with α-mPD-L1 were also cut into small pieces and digested at 37 °C with Collagenase I. Cell suspension was filtered through a 70 µm cell strainer and incubated with ACK Lysis solution (Gibco, Grand Island, USA). After a last filtration, the cell suspension was placed in a T75 flask at 37 °C, overnight. The day after, adherent cells were washed with PBS, trypsinized, collected and centrifuged. Cell pellets were stored at −80 °C until protein extraction and Western Blotting analyses, performed as described above.
Experiments involving animals have been were approved by the Italian Ministry of Health (Authorizations 213/2016 PR) and have been done in accordance with the applicable Italian laws (D.L.vo 26/14 and following amendments), the Institutional Animal Care and Use Committee of CEINGE and Allevamenti Plaisant SRL.