In cells, intracellular protein-protein interactions (PPIs) play vital roles. They facilitate the formation of protein-protein complexes that aid signal transduction and facilitate the binding of transcription factors to promoters and enhancers. Pharmacological strategies to inhibit intracellular PPIs have utilized small molecules with molecular weights less than 500 Da and biologics based on proteins with molecular weights greater than 5000 Da. Small molecules are efficient candidates in terms of cell permeability and inhibition of intracellular proteins [1, 2]. But their activity is hindered when a single mutation occurs at a target site, and cancer cells can easily acquire resistance against these drug candidates. Moreover, although the surface interaction between two proteins could be too large to be covered by small molecule candidates [3], there are currently several small molecule inhibitors, targeting the programmed cell death protein 1 (PD-1)/programmed death ligand 1 (PD-L1) interaction, undergoing clinical trials [4, 5]. On the other hand, large molecules such as antibodies can achieve high selectivity and affinity by binding to a large protein surface, but they face challenges when it comes to cell permeability [6]. Biotechnological advances leveraging aptamers and peptides represent a cutting-edge approach in refining the targeting PPIs.

Aptamers, short single-stranded DNA or RNA molecules [7, 8, 9], and peptides offer precise targeting capabilities due to their high specificity and affinity for proteins. By exploiting these biotechnological tools, researchers can design novel therapeutics capable of modulating the protein-based pathway with enhanced precision and efficacy. Furthermore, the versatility of aptamers and peptides allows for the development of multifunctional agents capable of synergistically targeting proteins while simultaneously delivering payloads such as drugs or imaging agents. In this review, we aim to provide a synthesis of the most recent advancements in biotechnological approaches employing aptamers and peptides for targeting Programmed Death Ligand 1 (PD-L1). These innovative strategies hold significant potential for not only enhancing personalized cancer diagnostics but also revolutionizing treatment strategies. By examining the latest progress in this field, we endeavor to shed light on the promising convergence of biotechnology and immunotherapy. Immunotherapy has become a significant research focus [10]. Presently, various approaches are employed in cancer treatment, including radiotherapy, chemotherapy, and surgery, each adhering to its own set of medication administration standards. However, these methods are constrained by their respective limitations. In contrast, immunotherapy stands out for its notable specificity, binding affinity, and low toxicity in tumor treatment. Significant research efforts have been dedicated to this burgeoning field.

A recent breakthrough in cancer immunotherapy is the immunologic checkpoint blockade [11], among which the programmed cell death protein 1 (PD-1) has been most actively studied [12]. PD-1 is expressed on the surface of activated T cells, B cells and myeloid cells [13, 14]. PD-1 has two natural ligands, PD-L1 and PD-L2 [15, 16]. Engagement of PD-1 with either ligand suppresses immune responses and promotes self-tolerance [15, 16, 17]. PD-L1 and PD-L2 only share 37% sequence identity, but have similar functions and expression profiles [18, 19]. PD-L1 is expressed in immune cells such as T and B cells, dendritic cells, and macrophages, as well as in many different tumor types [20]. PD-L2 expression has been reported to be more restricted to antigen-presenting cells [21]. The overexpression of PD-L1 and/or PD-L2 in cancer cells reduces the body’s immune responses, enabling cancer cells to evade killing mediated by T cells [22, 23]. PD-L1 is expressed more widely by tumor cells than PD-L2, and the blockade of the PD-1/PD-L1 interaction is more frequently targeted by therapeutic agents. The 2018 Nobel Prize in Medicine, to James P. Allison and Tasuku Honjo, acknowledged the research behind this technique [24]. To date, several therapeutic antibodies, targeting PD-1 or PD-L1, have been approved by the FDA, and a few others are currently undergoing clinical trials [25, 26].

While monoclonal antibodies have achieved significant success in clinical trials, it is important to acknowledge that the overall pooled response rates of anti-PD-1/PD-L1 antibodies have been reported to be below 25% [27, 28]. In additional, monoclonal antibodies are expensive with some other shortcomings, such as lower oral bioavailability and poor tumor penetration [29, 30]. Although small-molecule compounds have significant advantages in dealing with these problems compared to antibodies, it is very difficult to design small molecules targeting the PD-1/PD-L1 interaction [18, 31]. The main reason is, due to the large, flat, and hydrophobic areas of the PD-1/PD-L1 interaction surface [32], small molecules usually cannot offer large surface area to break the interaction. Until now, only very limited small molecules show moderate blockade activity of PD-1/PD-L1 binding [33, 34, 35]. On the other hand, nucleic acid aptamers and peptides, which are larger than small molecules but smaller than antibodies, are promising alternatives to target the PD-1/PD-L1 interaction [36]. Numerous endeavors have been undertaken thus far to develop aptamers or peptides as highly sought-after alternatives to antibodies, particularly as PD-1 or PD-L1 inhibitors. PD-1 is expressed in immune cells and PD-L1 is overexpressed in cancer cells; therefore, inhibitors targeting PD-1 may have side effects resulting in immune disorders, while PD-L1 inhibitors may reduce these problems [37, 38]. In addition, therapeutic PD-L1 antibodies have been reported to be more effective than PD-1 antibodies in blocking PD-1/PD-L1 signaling [39]. In this review, we will focus on discussing aptamers and peptides against PD-L1.

DNA and RNA aptamers, both classes of nucleic acid molecules, exhibit distinct characteristics that influence their applications in research and therapeutics. This review predominantly focuses on DNA aptamers, with only one RNA aptamer targeting PD-L1 discussed, highlighting the potential challenges in obtaining RNA aptamers. These challenges may stem from factors such as nuclease susceptibility and structural stability. Compared to RNA, DNA aptamers are inherently more stable in biological environments. The absence of the 2′-OH group in DNA makes it less susceptible to hydrolysis and self-cleavage. They are also typically easier and more cost-effective to synthesize, increasing their accessibility for researchers developing novel therapeutics or diagnostic tools. Conversely, RNA aptamers often exhibit higher binding affinity to targets due to the presence of 2′-hydroxyl groups, which can form additional hydrogen bonds with target molecules. This increased affinity can be advantageous in applications where strong target recognition is crucial. RNA aptamers also offer a broader range of target specificity owing to their ability to fold into complex secondary and tertiary structures, potentially allowing for recognition of a wider array of target molecules. The choice between DNA and RNA aptamers ultimately depends on the specific requirements of the intended application, with each class offering unique advantages and considerations. Notably, the lack of reliable 3D structure prediction for single-stranded DNA (ssDNA) led to the modeling of aptPDL-1 as an RNA aptamer using RNAcomposer [40], underscoring the practical differences between working with DNA and RNA aptamers. However, modern techniques are increasingly addressing these challenges, narrowing the gap between these two classes of aptamers.

The evolution of aptamer selection techniques has addressed several challenges in targeted therapeutics development. Traditional Systematic Evolution of Ligands by Exponential Enrichment (SELEX) methods often struggle with achieving high specificity and affinity, leading to the development of innovative variants. Modular-SELEX combines multiple SELEX types to overcome limitations in affinity, specificity, and compatibility [41]. X-Aptamer libraries incorporate modified nucleotides, enhancing chemical diversity and interaction robustness [42]. These approaches help address the fundamental challenge of aptamer degradation by nucleases in the body. To combat nuclease-mediated degradation and improve in vivo stability, researchers have incorporated non-canonical nucleotides into aptamers. For instance, 2’-fluoropyrimidine-RNA aptamers have been developed for triple-negative breast cancer (TNBC) cell recognition, potentially enhancing their stability in biological environments [43]. Similarly, the use of modified dU in X-Aptamer libraries not only improves binding affinity but also contributes to increased nuclease resistance [42]. Bispecific aptamers like Ap3-7c demonstrate how structural modifications can enhance both functionality and stability [44]. REase-SELEX introduces mutations that may contribute to improved aptamer stability and affinity [45]. The loss-gain cell-SELEX strategy focuses on membrane protein targeting, potentially reducing off-target interactions and improving specificity [46]. These advancements in SELEX techniques and aptamer modifications are crucial steps toward developing stable, highly specific aptamers capable of withstanding physiological conditions, thereby increasing their potential for diagnostic and therapeutic applications.

We first explore the landscape of anti-PD-L1 aptamers, detailing their discovery and discussing strategies for enhancing their efficacy, which underscores the promising potential of aptamer-based therapeutics in cancer immunotherapy. A novel DNA aptamer called aptPD-L1 was reported by Lai et al. [40] in 2016 for the first time against PD-L1 protein (Fig. 1, Ref. [40]). The aptamer showed promising results in in vitro lymphocyte proliferation and suppressed in vivo tumor growth without noticeable liver or renal toxicity. The study findings demonstrated the effectiveness of restoring T-cell function and altering the microenvironment of tumors. The aptamer is 45 nucleotides long, which specifically binds to the target PD-L1 with a dissociation constant (Kd) of 4.7 nmol/L compared to bovine serum albumin (BSA) and blocks 58% of the interaction between PD-1/PD-L1. The aptPD-L1 sequence was 5^′-ACG GGC CAC ATC AAC TCA TTG ATA GAC AAT GCG TCC ACT GCC CGT-3^′ and since then it has been used in different research [40].

A year later, in 2017, the protein SELEX method using magnetic beads was performed, and two possible aptamers (Apt 5 and Apt 33) targeted PD-L1 was reported [47]. Of these, Apt 5 showed high affinity and specificity with a Kd of 64.77 nM compared to the other candidate and was examined further. The selected aptamer was labelled with ATTO 647N to determine its dissociation constant. Flow cytometry data showed a better binding affinity for Apt 5 with a fluorescence log intensity of 3.89 $\times$ 10⁵. Likewise, this aptamer was subjected to cell selectivity and internalization assays, and the results indicated that it could be used as a promising agent for diagnostic and therapeutic delivery. The reported sequences for Apt 5 was 5^′-GCT GTG TGA CTC CTG CAA GAC GGA CCA GCC TTG CCG CAA GAC GGA CCA GGG ATT CAA ACG AGC AGC TGT ATC TTG TCT CC-3^′. The selected aptamer detected PD-L1 expressing cells as low as 10 cells mL^-1 [47].

Similarly, another anti-PD-L1 aptamer has been reported to interfere with the interaction between PD-1/PD-L1 and its antitumor effect. Cell SELEX was performed using engineered PD-L1 expressing cells on a target. This work resulted in a highly specific binding aptamer PL1 with Kd value of 95.73 $\pm$ 14.27 nM which showed the inhibition of the interacting target protein. P1 slows the growth of CT26 colon cancer cells. Moreover, it has the ability to boost the growth of T-cells that have been retarded by PD-1/PD-L1 interaction. It helped to rescue the T-cell by increasing the production of interferon-gamma (IFN-_Ɣ) The sequence (5^′-ATA CCA GCT TAT TCA ATT GTA GAG TAT AAA AAG AGT GAT GAT CTT TTG TAG GTT TT TTA GAT AGT AAG TGC AAT CT-3^′) was designated PL1, which did not show binding affinity with Chinese hamster ovary cell line K1(CHO-K1) cells, which were used as control cells in the experiments. The specificity of the aptamer was determined using other cells, such as U87, 7880, and HCT 116, which showed no recognition except for the targeted cell. To check the inhibition efficacy of PL1, in vivo and in vitro studies were performed. The surface plasmon resonance (SPR) result showed the high binding of PL1 with an inhibition of 82% compared with other random sequences [48].

The aptamer used in the homogeneous, low-volume, efficient, and sensitive exosomal programmed death-ligand 1 (HOLMES-Exo_PD-L1) technique is MJ5C [49], a short single-stranded DNA molecule engineered specifically to efficiently target PD-L1. MJ5C exhibits significant affinity for PD-L1 as evidenced by its binding Kd of 91 $\pm$ 12 nm. The 33-nucleotide sequence of MJ5C (5^′-TAC AGG TTC TGG GGG GTG GGT GGG GAA CCT GTT-3^′) was refined from the original MJ5 aptamer by a selection procedure that considered how well it binds to the PD-L1 cancer cell surface. Validation studies have demonstrated that MJ5C is an effective technique for identifying exosomal PD-L1 in cancer patients, offering superior selectivity and higher binding efficiency to native PD-L1 compared to antibodies, as shown in Fig. 2 (Ref. [49]). The findings suggest that glycosylation of PD-L1 impairs antibody recognition, whereas the PD-L1 MJ5C aptamer, being approximately ten times smaller, is less affected by steric hindrance [49].

Through the use of Modular-SELEX, clon-3 aptamer was identified [41]. High compatibility and specificity was demonstrated using clon-3 in identifying PD-L1 expression in tumor samples. Aptamer dissociation constant (Kd) values ​​were obtained using the waveguide sensor at 70.1 $\pm$ 14.2 nM and the Nano-Affi assay at 8.6 $\pm$ 1.9 nM. In addition, the Kd of clon-3 against BCPAP cells expressing high levels of PD-L1 was 95.9 $\pm$ 29.8 nM, which was similar to the Kd of another anti-PD-L1 aptamer identified by MB-SELEX. These findings demonstrate the high binding affinity of clon-3 having sequence 5^′-CCC TCC TCC TAA CTG TTC CTA CGA AAC GAG CTT ATG CGT AAT GAT GAC TGT CGT AGT TCG-3^′ for PD-L1, reinforcing its promise for tumor tissue imaging and cancer cell detection [41].

In another research, aptamers A6 and B10 were studied with their high affinity or specificity against PD-1 and PD-L1, respectively [50]. For PD-1, aptamer A6 had a Kd value of 47.84 $\pm$ 24.78 nM and aptamer B10 had a Kd value of 59.72 $\pm$ 15.87 nM for PD-L1. Further validation research, including EMSA (Electrophoretic Mobility Shift Assay) and dot blot assays, confirmed the high affinity and specificity of A6 for PD-1 and B10 for PD-L1. Cloning and sequencing were used to produce ten aptamer sequences for both rhPD-L1 and rhPD-1 in the screening phase. Dot blot, electrophoretic mobility shift, and gold nanoparticle-based colorimetric assays were used to evaluate the specificity and affinity of potential aptamers. The B10 aptamer having sequencing 5^′-AAC AAC GTA TAA CAA TGC CCA CGT CAC CAG AGT ACT ATG G-3^′ was shown to be most effective for identifying PD-L1, highlighting the potential of this protein for tailored binding in anticancer applications [50].

When non-natural or modified nucleotides are incorporated into the aptamer sequence, it allows for improved binding qualities and a wider range of functions than conventional aptamers. This is referred to as “Xeno” in X-Aptamers. X-Aptamers XA-PD1-78 and XA-PD-L1-82 were selected using a bead-based library of X-Aptamers with various changes, including amino acid functional groups linked to dU at position 5. In pancreatic tumor tissue, XA-PD-L1-82 showed binding capacity comparable to PD-L1 antibodies, suggesting its potential for therapeutic use. Strong binding affinities were observed for the binding of XA-PD1-78 and XA-PDL1-82 (Kd = 56.72 nM and 20.18 nM, respectively) to cells expressing PD-L1. These X-Aptamers provide a synthetic replacement for antibodies in PD-L1 research, diagnostics, and possible therapeutic treatment [42].

As shown in Fig. 3 (Ref. [51]), researchers discovered two threose nucleic acid (TNA) aptamers, N5 and S42, that effectively inhibited the PD-1/PD-L1 interaction through an in vitro selection procedure. These TNA aptamers showed more than 50% inhibition of PD-1/PD-L1 interaction in enzyme-linked immunosorbent assay (ELISA) experiments (Fig. 3) and their binding affinities for PD-L1 protein were approximately 400 nM. Crucially, TNA aptamers also showed their efficacy in a more physiologically realistic setting by disrupting the PD-1/PD-L1 junction on the cell surface. Moreover, TNA aptamer N5 with reported truncated sequencing 5^′-GAT TGA GTA GAT AGT GGT TCT GTA CGT AGT GAA AGA GTG G-3^′ selectively accumulated at the tumor site and dramatically suppressed tumor growth in mouse tumor xenograft experiments, all without apparent damage. These aptamers differ from normal DNA or RNA aptamers due to their TNA composition, which confers increased resistance to nucleases and may increase their therapeutic value as immune checkpoint inhibitors for cancer immunotherapy [51].

In another study, DNA aptamer Ap3 was found and characterized. It binds preferentially to PD-L1 on tumor cells, with a Kd value of 169.0 nM for the protein and 204.2 nM for cells expressing PD-L1, respectively. The research team then combined Ap3 with an anti-PD1 aptamer to create a bispecific aptamer called Ap3-7c. They also invented a “recognize then conjugate” method based on click chemistry to produce a covalently attached bispecific aptamer called D-Ap3-7c. This innovative approach showed potent anticancer effects in vivo, demonstrating the unique ability of aptamers to target multiple immunological checkpoints simultaneously and enhance immunotherapy [44].

Using a novel SELEX approach called REase-SELEX, the type II restriction enzyme Alu I was critical in aptamer isolation. Alu I cleaved the non-binding material and caused massive variation in the DNA library, greatly expanding its spectrum. As demonstrated by many experiments, aptamer 8–60 showed high affinity (Kd = 1.4 nM) and specificity towards PD-L1. Notably, 8–60 outperformed a previously identified aptamer in reliably detecting PD-L1 expression levels in tissue sections from malignant melanoma, non-small cell lung cancer, and normal human tonsils. They concluded that REase-SELEX can isolate highly specific aptamers by producing targeted mutations during the selection process [45].

Recently, using a novel cell loss-gain-SELEX approach, the DNA aptamer (XQ-P3) was identified, which binds to PD-L1 with high affinity (Kd ~15 nM) and specificity. In triple-negative breast cancer cells that overexpress PD-L1, the XQ-P3 aptamer was able to improve the cellular uptake and cytotoxicity of the conjugated chemotherapy drug paclitaxel, as well as suppress the PD-1/PD-L1 interaction and restore T cell activity. The researchers used a variety of methodologies such as pull-down assays, competitive binding, and cell-based assays to verify the specificity of XQ-P3 for PD-L1 with a sequence of 5^′-ACC GAC CGT GCT GGA CTC ATC TCG CTT TTT TCA CGG TCC ACA CTA CTA TGA GCG AGC CTG GCG-3^′. These results identified the potential of the PD-L1 aptamer for immunomodulation and targeted treatment of triple-negative breast cancer [46].

In this section, we have thoroughly explored the field of PD-L1 aptamers, discussing their discovery and strategies for enhancing their effectiveness through conjugation with other aptamers or alternative approaches. We emphasize the promising potential of aptamer-based therapeutics in cancer immunotherapy. Table 1 (Ref. [40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52]) provides a summary of the general information regarding these anti-PD-L1 aptamers.

In addressing the limitations of small molecules and biologicals, peptides that interfere with PPIs with a molecular weight ranging between 500 and 5000 have been developed [53]. They are easy to synthesize, have lower immunogenicity, better pharmacokinetics, and lack Fc-receptor-related side effects [54, 55, 56, 57]. Due to this unique property, they can act as a binding bridge between monoclonal antibodies and small molecules. Moreover, the peptides-based inhibitor can be easily modified for better stability and to penetrate tumors and effectively block the interaction between PD-1 and PD-L1 in both near and distant areas of the tumor’s blood vessels. The most common modification involves macrocyclization and stapling of peptides. Macrocylization of peptide enhances the rigidity in structure which help in improving binding specificity and resistance in proteolysis. They are resistant to degradation, target protein-protein interactions, and show promise in therapeutic development for disease like cancer. Recent advancements have expanded their potential in drug discovery [58]. On the other hand, stapling of peptides makes it more constrained, which supports further stability and bioavailability enhancing their efficacy as treatments. This technique involves reinforcing peptides by introducing a chemical “staple”, thereby making them more resistant to degradation within the body. This method is instrumental in developing drugs that specifically target protein interactions and has demonstrated potential in inhibiting disease-associated proteins [59]. Apart from macrocyclic peptide and stapled peptide, there is another class of peptide molecule known as miniproteins. These molecules are a diverse group of small-sized protein frameworks (1–10 kDa) known for their stability and effectiveness in drug-related functions. These miniproteins are mainly derived from natural sources and have been widely used in pharmaceutical development. Despite their different structures and functions, the methods for their application show more similarities among themselves than with more commonly used modalities such as antibodies or small molecules [60]. Peptide-based therapeutics also provide options like nano-scale drug delivery with easy conjugation to a targeting ligand or encapsulated system. Because of these advantages, peptide-based inhibitors have a future as remarkable candidates for cancer immunotherapy [61, 62]. Extensive studies have been carried out in various scientific research institutions to inhibit PD-L1 by using various display technologies like phage display, bacterial display, and mRNA display and by computational design peptide mimetic. These techniques are used to study protein interactions and identify peptides with high affinity for specific targets. Phage display involves expressing peptides on the surface of bacteriophages, which allows for easy manipulation and amplification of the peptide library [63]. Similarly, bacterial display involves expressing peptides on the surface of bacteria, often E. coli, which offers advantages in terms of speed and cost due to the fast growth and simple culture conditions of bacteria [64]. Furthermore, mRNA display links the peptide to its encoding mRNA via a covalent bond, allowing for the direct selection of peptides with desired properties from vast libraries and facilitating rapid evolution of high-affinity binders [65]. On the other hand, the Computational design of peptide mimetics leverages advanced algorithms and molecular modeling to design peptides that mimic the structure and function of natural proteins, providing a powerful approach to creating novel therapeutics and enhancing drug discovery processes [66]. The peptides discovered from these technologies are synthesized and purified by different way. One of common way is Solid-phase peptide synthesis (SPPS) that allows for the step-by-step addition of amino acids on a solid resin support, making purification easier and increasing yield [67]. Liquid-phase peptide synthesis (LPPS) is another technique that is often used for longer peptides because it is easy to handle and purify intermediates [68]. Recombinant DNA technology is widely used for large-scale production, where peptides are expressed in microbial systems such as E. coli or yeast [69]. Enzymatic synthesis, which uses proteases and peptidases, offers a biocatalytic route to peptide production with high specificity and mild reaction conditions [70]. Each method has its own advantages, and the choice depends on factors such as peptide length, complexity, and application requirements. Herein, we review the recent advances on peptide-based inhibitors targeting PD-L1.

To date, many peptide-based PD-L1 inhibitors have been reported. In 2014, Bristol-Myers Squibb [71] introduced a significant category of macrocyclic peptides derived from peptidomimetics, aiming at the PD-1/PD-L1 pathway. They disclosed three cyclic peptides Bristol-Myers Squibb (BMS) 57, BMS 71, and BMS 99—comprising 15, 14, and 13 amino acids, respectively. Magiera-Mularz et al. [72] conducted a study on the interactions between these peptides and PD-L1, elucidating the binding mechanisms along with certain aspects of their biological activity. The reported IC₅₀ values for BMS-57 and BMS-71 were found to be 9 nM and 7 nM, respectively. Various techniques such as nuclear magnetic resonance (NMR)titration, differential scanning fluorimetry (DSF), and cell-based PD-1/PD-L1 blockade bioassay were employed to corroborate the affinities of BMS-57 and BMS-71 towards PD-L1 [72]. NMR titration analysis demonstrated that these two macrocyclic peptides lack activity towards PD-1, whereas they exhibited binding affinities with PD-L1, with Kd values significantly below 0.1 µM. These findings were corroborated by DSF tests and further validated through TCR experiments. Additionally, cell-line experiments revealed EC50 values of 566 nM and 293 nM for BMS-57 and BMS-71, respectively [73, 74, 75].

In a separate study, Chang et al. [76] identified the peptide sequence FPNWSLRPMNQM through phage display peptide screening, which exhibited a strong binding affinity for the PD-L1 protein. Subsequent investigations revealed that the peptide degraded rapidly in blood plasma. To address this stability issue, they developed an enzymolysis-resistant D-peptide by combining techniques involving chemical synthesis of D-proteins and mirror-image phage display (Fig. 4, Ref. [76]). The resulting peptide sequence NYSKPTDRQYHF (Fig. 5A) demonstrated a binding affinity of 0.51 µM and significant inhibition of tumor growth in a CT26 transplantation model, with no observed loss of body mass.

In 2018, Li et al. [77] reported on a 22 amino acid residue peptide named targeting PD-L1 peptide (TPP-1), which targeted PD-L1. Utilizing the bacterial surface display method, they identified the peptide with the amino acid sequence SGQYASYHCWCWRDPGRSGGSK (Fig. 5B), possessing a binding affinity of 94 nM. The interference of TPP-1 peptide with the interaction between PD-1/PD-L1 was confirmed through a T-cell activation assay and mixed lymphocyte reaction. To evaluate TPP-1’s inhibitory effect on tumor growth in vivo, a xenograft mouse model was established using H460 cells. Mice treated with TPP-1 or PD-L1 antibody exhibited a 56% or 71% reduction in tumor mass growth rate, respectively, compared to those treated with a control peptide. These findings suggest that TPP-1 has the potential to inhibit or at least decelerate tumor growth [77].

In 2019, Liu et al. [78] developed a novel phage display strategy to identify a peptide capable of specifically binding to hPD-L1 residues responsible for the PD-1/PD-L1 interaction. This approach was further validated through docking studies. They identified four peptides (CLP001, CLP002, CLP003, CLP004), among which CLP002, with the sequence WHRSYYTWNLNT (Fig. 5C), exhibited the highest blocking efficiency (85%) and a binding affinity of 366 nM for PD-1/PD-L1 interaction. Additionally, CLP002 demonstrated an IC₅₀ of 1.43 µM against DU-145, a prostate cancer cell line, and 2.22 µM against hPD-L1. Further investigations revealed that CLP002 was capable of restoring T-cell proliferation and preventing apoptosis during co-culture with cancer cells. Moreover, it displayed superior tumor penetration in a 3D tumor spheroid model compared to Cy5-labeled anti-PD-L1 antibody. Furthermore, CLP002 inhibited the growth of CT26 colorectal tumors and increased the survival of tumor-bearing mice [78].

Next, Gurung et al. [79] identified two PD-L1 binding peptides, PD-L1 pep-1 and PD-L1 pep-2, using a T7 phage library containing CX7C peptide. Bio-panning was conducted on transfected HEK293T cells, with counter-selection performed on non-transfected HEK293T cells. The peptide sequences CLQKTPKQC and CVRARTR (Fig. 5D,E) exhibited selective binding affinities of 373 nM and 281 nM, respectively, toward high PD-L1-expressing cells such as MDA-MB231 and CT26 tumor cells, over low PD-L1-expressing cells like MCF7 tumor cells. This binding was further augmented by treating cells with IFN-$\gamma$, a potent inducer of PD-L1 expression. The cellular binding of pep-1 and pep-2 was validated through pull-down assays of PD-L1 using biotin-labeled peptides, PD-L1 gene expression knockdown using siRNA, and competition assays with an anti-PD-L1 antibody. In addition to inducing T-cell reinvigoration through PD-L1 blocking activity, these peptides demonstrated targeted delivery of chemotherapeutic drugs to tumors with high PD-L1 expression. PD-L1-targeted liposomal doxorubicin showed more efficient inhibition of tumor growth compared to naive liposomal doxorubicin. This enhancement was achieved with lower concentrations of PD-L1 pep-2 and doxorubicin, compared to combination therapy with PD-L1 pep-2 and free doxorubicin [79].

In 2021, Yin et al. [80] utilized computational design to create a miniprotein mimetic aimed at blocking the PD-1/PD-L1 interaction by imitating the interface of an optimized PD-1. Initially, attempts to mimic the native structure of PD-1 resulted in the MNPD-1 peptide, which displayed no activity despite mimicking nine of the 22 hotspot residues from the interaction surface. However, with the optimized PD-1, they identified the mimetic peptide Mimetic of Optimized PD-1 (MOPD-1) (IQIREYKRCGQDEERVRRECKERGERQNCHYVIHKEGNCYVCGIICL), which exhibited a binding affinity of 0.3 µM for PD-L1. Remarkably, the designed peptide inhibited the interaction with an IC₅₀ of 1–2 µM, irrespective of whether PD-1 or PD-L1 was immobilized. Further characterization using NMR spectroscopy and X-ray crystallography revealed that the binding residues from the modified PD-1 were crucial for the bioactivity of MOPD-1. Moreover, the mimetic peptides were found to be highly stable, with serum half-lives exceeding 24 hours without requiring any modifications such as cyclization. In vivo studies demonstrated that MOPD-1 suppressed tumor growth by revitalizing the immune system in a CD8⁺ T cell-dependent manner [80].

Recently, Fetse et al. [81] employed a side chain macrocyclization strategy with two cysteine residues scanning on the anti-PD-L1 linear peptide CLP-002 to discover a cyclic peptide with improved inhibition and binding affinity. Among 22 scans, they found that peptides C7 (CHRSYYTWNLNC, Fig. 5F) and C12 (CHRSYYCWNLNT, Fig. 5G) exhibited superior binding affinity compared to the unmodified CLP-002 peptide. Notably, the C7 peptide demonstrated an IC₅₀ of 180 nM, which was 34 times better than the parent CLP-002 peptide’s IC₅₀ of 6073 nM against mouse PD-1/PD-L1 interaction. The efficacy of C7 and C12 peptides was further supported by serum stability experiments. While the CLP-002 linear peptide displayed a half-life of 18.8 minutes and was nearly completely degraded after 3 hours in 50% human serum, the cyclic peptides C7 and C12 exhibited improved serum stability, with half-lives of 38 and 70 minutes, respectively. Among these, the C12 peptide showed the best serum stability, with approximately 25% of the peptide remaining intact after 3 hours. Furthermore, both cyclic peptides demonstrated superior in vivo anti-tumor efficacy even at a dose of 0.5 mg/kg compared to the parent peptide CLP-002 at 2 mg/kg [81].

Kamalinia et al. [82], in 2020, engineered an 18-residue linear peptide called SPAM using mRNA display, which binds to human PD-L1. The SPAM peptide has sequence MPIFLDHILNKFWILHYA (Fig. 5H), which shows different characteristics compared to known PD-L1 binding peptides and mAbs. It exhibits high selectivity for human PD-L1, with dissociation constants of 119 and 67 nM for unglycosylated and glycosylated human PD-L1, respectively. The SPAM peptide does not significantly bind to mouse PD-L1 or human PD-L2. Additionally, competition binding assays suggest that the SPAM peptide’s binding site overlaps with the binding site of PD-1 and therapeutic anti-PD-L1 antibodies. The peptide shows serum stability for up to 5 hours with no change in peptide retention time in 25% human serum [82].

In 2023, Shi et al. [83] discovered a peptide sequence that interacts with the protein acyltransferase ZDHHC3, responsible for modifying the PD-L1 protein. Utilizing artificial intelligence, they successfully pinpointed the precise peptide sequences to target ZDHHC3. Subsequently, these sequences were modified through strategic stapling at optimal positions to enhance stability and effectiveness. This stapling technique significantly improved the peptides’ stability, binding capacity to the target, and ability to penetrate cells. The resulting stapled peptide was conjugated with linker GSGS, and E3 ligase peptide ALAPYIP, named SP-PROTAC, was employed as a peptide-based PROTAC, effectively reducing PD-L1 expression to less than 50% at 0.1 µM in both C33A and HeLa cervical cancer cell lines. Furthermore, the expression of ZDHHC3 decreased in a dose and time-dependent manner. In a C33A and T cell co-culture model, the use of the peptide led to a dose-dependent increase in IFN-$\gamma$ and Tumor Necrosis Factor Alpha (TNF-$\alpha$) release through PD-L1 degradation [83].

BMS986189, a macrocyclic peptide that entered the clinical trial, provided favorable safety and pharmacokinetic data. The results from a Phase I clinical trial (ClinicalTrials.gov NCT02739373; https://clinicaltrials.gov/ct2/show/NCT02739373) completed in December 2016 led to the development of an analog of the original candidate, BMS986189, for which a new Phase I clinical trial is ongoing in 2022 (ISRCTN Registry ISRCTN17572332, https://www.isrctn.com/ISRCTN17572332). Inspired from BMS986189, Rodriguez et al. [84] did structural and biological characterization of pAC65 macrocyclic peptide and found its potency equivalent to the current FDA-approved mAbs and BMS986189. The verification of its potency is based on various assay. In Antagonist Induced Dissociation Assay (AIDA)-NMR experiments, they found that pAC57 is capable of interfering with both the PD-1/PD-L1 and PD-L1/CD80 complexes formation. To further support its effectiveness, they did a commercially available Homogeneous Time Resolved Fluorescence (HTRF) assay which shows affinity of pAC65 for PD-L1 and dissociate PD-1/PD-L1 complex with an IC₅₀ value of 1.80 $\pm$ 0.16 nM. In the structural characterization, it was observed that the peptide ring has an ellipse-like shape with a bulge located above the central region. It binds parallel to the plane of the $\beta$-sheet of the PD-L1 protein. The backbone of the macrocyclic ring extends from 65Ala-NH25 to 65TrpNAc10, while the bulge is formed by 65Tyr1, 65NMeNle12, 65Leu13, and 65Scc14. The bioactivity and lack of toxicity of pAC65, along with its similarity to BMS986189, make it an ideal candidate for preclinical testing. This could confirm its potential, safety, and tolerability. The discovery of a macrocyclic peptide with similar activity to mAb could lead to the development of macrocyclic peptides suitable for administration via routes like intranasal or pulmonary, pending supportive preclinical data [84].

In this section, we have extensively explored the realm of PD-L1 peptides, detailing their discovery and discussing strategies for enhancing their efficacy through conjugation with other molecules. We underscore the promising prospects of peptide-based therapeutics in cancer immunotherapy. Fig. 5 depicts the structures of several anti-PD-L1 peptides elucidated in this review. Moreover, Table 2 (Ref. [77, 78, 79, 80, 81, 82, 83]) summarizes the general information pertaining to these anti-PD-L1 peptides.

Both dimerization and multiple presentations serve as valuable tools in aptamer or peptide development, providing avenues to enhance binding properties and potentially incorporate supplementary functionalities. The specific approach chosen depends on the desired outcome and the characteristics of the target molecule. In sections 4.1 and 4.2, we will examine recent developments in dimerization of anti-PD-L1 aptamers and peptides, respectively. In the following section 4.3, we will delve into their multiple presentations.

In conclusion, the current scientific literature lacks comprehensive reviews that explore both aptamers and peptides as potential tools for targeting PD-L1. Addressing this gap is crucial to consolidate recent advancements in biotechnological approaches aimed at PD-L1 targeting. Our review aims to fill this void by synthesizing available data and discussing the potential implications of aptamer- and peptide-based strategies. Such insights would offer valuable guidance for researchers and clinicians striving to optimize PD-L1-targeted therapies, enhancing efficacy and precision in cancer treatment.

HT, SA, and XT conducted a comprehensive literature review, synthesized the findings through the creation of illustrative charts, and developed the initial conceptual framework for the manuscript. All authors contributed critically to the subsequent manuscript drafts, incorporating feedback from the review process to produce the final version. All authors read and approved the final manuscript. All authors have participated sufficiently in the work and agreed to be accountable for all aspects of the work.

Not applicable.

Not applicable.

The authors thank Bowling Green State University for Startup grant (33900025) and the Building Strength Grant Programs to XT (18700).

The authors declare no conflict of interest.