There are several treatment options available against cancer, typically including surgery, radiotherapy, chemotherapy, and hormonal treatments, which are often associated with adverse side effects. Besides other recently developed, more specific and individual therapies such as immunotherapies and cell therapies, oncolytic viruses (OVs) such as oncolytic adenoviruses (OAds) represent an attractive alternative treatment option with high potential. OAds selectively infect and kill cancer cells, and they are explored in clinical settings as a form of cancer immunotherapy, either alone or in combination with conventional treatments or immune checkpoint inhibitors.

Adenoviruses (Ads) are nonenveloped viruses with an icosahedral capsid, which is 70–90 nm in size 1, 2. There are 116 identified human Ad types [3] divided into the species A to G and even more nonhuman Ads isolated from numerous animals. Species D comprises the largest group of human Ad types, whereas species E, F, and G each contain one or two Ad types 1, 4. Species B Ad types contain 20 and species C 7 Ad types [3]. Ads contain a linear, double-stranded DNA genome that is typically around 26 to 48 kb in size, depending on the specific Ad type. It is organized into a linear structure flanked by inverted terminal repeats (ITRs) and encodes approximately 30 to 40 open reading frames, which are responsible for various functions in the viral life cycle, including replication, transcription, and assembly. The genome is divided into several key regions such as the early transcription units E1 to E4 and the late transcription units L1 to L5 1, 5, 6. The icosahedral capsid is composed of three main proteins, including hexon, penton, and fiber. In addition, the fibers arise from the 12 vertices of the icosahedron, whereas the penton base is located at the base of each fiber [1].

Numerous clinical trials were performed with OAds solely based on human adenovirus type 5 (HAdV-C5) as backbone. Prominent examples are Oncorine (H101) containing an E1B-55k deletion and a complete deletion of the E3 region 7, 8, ONYX-015 containing an 827 bp deletion in the E1B region and a point mutation [9], and HAdV-C5 OAds with an E1A protein containing a deletion of amino acids 120 to 127 (Δ24) that selectively targets cells with abnormal Rb control [10]. Advances in genetic engineering then enabled further advancements in OAd development. This included the insertion of tumor-specific promoters or immune-stimulatory transgenes into the Ad genome to further boost the anticancer effects. However, although 116 human Ads and even more nonhuman Ads were identified, representing a large natural diversity for genome engineering and translational approaches, there are limited studies for non-HAdV-C5–based OVs.

One of the major reasons why there are only limited studies for non-HAdV-C5–based OAds is the lack of fast and efficient options to clone and subsequently arbitrarily genetically modify captured non-HAdV-C5 Ad genomes. To get access to complete Ad genomes as virus or vector, different strategies were utilized, including traditional methods such as molecular cloning, cosmid-based methods [11], and homologous recombination (HR) in bacteria [12] and in eukaryotic cells 11, 13, 14 (reviewed in Ref. [15]). Recently, more advanced methods were explored to capture complete Ad genomes and to genetically modify them such as module-based adenoviral genome assembly and advanced HR methods in bacteria HRB.

Module-based adenoviral genome generation, including capture of complete Ad genomes can be based on traditional cloning [16] or by advanced cloning and assembly methods to not generate scars (Gibson Assembly, NEBuilder HiFi DNA Assembly Cloning, In-Fusion cloning). Using the latter method, the different modules, containing parts of the Ad genome with short homology arms (HAs), are amplified by polymerase chain reaction (PCR) and subsequently assembled 17, 18 (Figure 1a). In another assembly strategy, segments of the Ad genome were individually amplified and cloned into plasmids as fragment. Following restriction enzyme digest or PCR amplification, Ad genome DNA fragments were directly added to the assembly reaction (Figure 1a). After assembly, the linearized complete Ad genomes are transfected to packaging cells 19, 20•.

Besides modern assembly methods, HRB has become an alternative method of choice when it comes to larger gene insertions or modifying a variety of different Ad types. Previous recombineering approaches include the Red recombination system of bacteriophage λ in E. coli strain SW102 capturing Ad genomes in bacterial artificial chromosomes using galK-positive/negative selection and linear-circular homologous recombination [21]. Because of the broad variety of different Ad types to be vectorized and modified, a high-throughput strategy based on HRB was invented, which utilizes the RecET recombinase system 22, 23, 24 (Figure 1b). In this method, the vectorization of the Ad genome starts with four PCR products, comprising the bacterial backbone, the 5´ ITR, a counter-selection marker cassette (SM) and the 3´ ITR. These four PCR products are flanked with HAs on each site to be recombined to one plasmid in E. coli strain GBRedGyrA462 with an L-arabinose-inducible recombinase activity. The SM contains an ampicillin resistance for positive selection and CcdB encoded from the CcdA/CcdB toxin–antitoxin system for counter-selection. In normal E. coli, the CcdB toxin would bind to the DNA-Gyrase complex, preventing resealing of the DNA during replication and therefore would lead to cell death. Due to the mutation in this specific E. coli strain, the CcdB toxin cannot target the gyrase A (GyrA) subunit of the DNA gyrase within the bacteria. Therefore, during counter-selection, the plasmids inheriting the SM can survive [25]. The shuttle plasmid is then linearized with restriction digest, and the Ad genome is inserted by linear–linear homologous recombination (LLHR) in E. coli strain GB05-dir, which also has an L-arabinose-inducible recombinase activity. To simplify this method, the right and left ITRs from the Ad genome to be vectorized were introduced as HAs in a linear PCR product containing the bacterial backbone from a plasmid containing the counter-selection marker. Together with the linear Ad genome, the LLHR takes place in a special E. coli strain, called GB05-dir with L-arabinose induced recombinase activity 22, 23, 24, 26, 27. After the discovery of recombineering and the RecET system, more than 40 different Ad types were vectorized to date 24, 28.

There are numerous OAds based on HAdV-C5 backbone with common modification sites, including the E1 and E3 transcription units, the fiber- and hexon-coding regions. A schematic overview of these modification sites is shown in Figure 2. Before completely switching the Ad type of OAds from HAdV-C5 to another Ad type, chimeric Ad vectors with a HAdV-C5 backbone were developed containing fibers or hexon regions (hypervariable regions [HVRs]) from alternative Ad types. These chimeric OAds and their exploration in preclinical and clinical studies are summarized in Table 1A and B.

Starting in 1996, OAds comprising chimeras of different Ad types were generated mainly to overcome the hurdle of Ad5 using CAR as its main receptor, which is often downregulated on tumor cells. Here, the fiber of Ad5 was modified to carry the knob of HAdV-B3 29, 30, 31, 32, HAdV-B34 [33], HAdV-B35 32, 34, 35, 36, HAdV-D37 32, 37, or HAdV-D49 [38]. The second strategy includes incorporation of arginine-glycine-aspartic acid (RGD) peptides in the fiber knobs of Ad5, which allows the virus to enter cells using αvβ3 or αvβ5 integrins 10, 39. Note that clinical trials applying RGD-modified OAd have shown promising results in glioma patients [40]. Recently, this virus was demonstrated to escape anti-HAdV-C5 neutralizing antibodies by exchanging the HVRs of the hexon with HVRs from the rare Ad type HAdV-D43 [41]. In another study, Ad-directed evolution by random peptide display into the fiber knob of HAdV-C5 was used to select improved OAds [20]. These OAds with the hTERT promoter for restricted replication in telomerase-expressing cell lines showed enhanced lysis in various tumor cells. Another study explored fiber modification in HAdV-C5 by insertion of the epitope of programmed cell death protein 1 (PD-1) (70-77 amino acids) at the HI-loop of the HAdV-C5 fiber. This OAd displayed a significant improvement in viral transduction efficiency in Programmed Cell Death-Ligand 1-positive cancer cells [42].

Fiber-modified viruses were also explored in clinical studies. Fiber-modified viruses Ad5/3-∆24-GM-CSF (ONCOS-102), Ad5/3-E2F-Δ24-GMCSF (CGTG-602), and Ad5/3-E2F-∆24-hTNFa-IRES-hIL2 (TILT-123) represent chimeric OAds with a HAdV-C5 backbone. ONCOS-102 was initially analyzed in a dose escalation study assessing safety, efficacy, and immunological end points after local treatment [43]. The results revealed that this treatment resulted in the infiltration of T-cells to tumors and upregulation of PD-L1. In addition, ONCOS-102 was explored in advanced melanoma, which were anti-PD-1 resistant [44]. Therefore, in this clinical trial, ONCOS-102 was combined with pembrolizumab, an antibody blocking the PD-1 receptor, which also promoted T-cell infiltration into the tumor side. In another clinical study, ONCOS-102 was used to treat patients with malignant pleural mesothelioma in combination with pemetrexed and cisplatin/carboplatin [45].

The chimeric OAd CGTG-602 [46] encoding the granulocyte-macrophage-colony-stimulating factor (GM-CSF) was investigated in 13 patients with solid tumors refractory to standard therapies. It was shown in this study that OAd treatment results in infiltrating T-cells when analyzing biopsies of the patients. Another interesting OAd is TILT-123-encoding tumor necrosis factor alpha (TNFα) and interleukin-2 (IL-2). This chimeric OAd, based on HAdV-C5 and carrying a HAdV-B3 fiber, was analyzed in various clinical trials 47, 48, 49•. One recent study applying TILT-123 included treatment of platinum-resistant refractory ovarian cancer patients in combination with pembrolizumab. Here, safety was assessed and results are encouraging to further explore this OAd in combination with other established anticancer therapies, such as checkpoint inhibitor anticancer drugs.

In theory, all investigated strategies applied for HAdV-C5 can also be used for OAds based on any other human Ad type. Therefore, it may be attractive to directly transfer genetic modifications used for HAdV-C5 to the alternative Ads derived from other human Ad species. Advantages of OAds derived from alternative Ad types are potentially (i) a lower seroprevalence in the human population, (ii) uptake into the tumor cell of the OAd mediated by an alternative non-CAR-dependent mechanism, and (iii) replication efficiencies of the explored novel Ad type may be cell type–specific and more robust in the explored tumor cell. However, the number of OAds for which a complete Ad type switch was performed is rather low. For instance, species A, E, F, G-derived Ads, were not explored as OAds until today. Therefore, in this review, we focus on species B, C, D OAds, and a summary of these studied OAds in preclinical and clinical studies is provided in Table 2A and B.

In 1985, it was shown for the first time that HAdV-C2 with a deletion in E1A shows enhanced oncolysis [50]. More than three decades later in 2011, the first species B-derived OAd completely based on HAdV-B3 was described [51]. As receptor for cell entry HAdV-B3 utilizes Desmoglein 2 [52] and the OAd Ad3-hTERT-E1A was developed using the hTERT promoter for tumor-specific replication. In 2017, the OAd Ad3-hTERT-CMV-hCD40L [53] for tumor cell–specific replication was introduced. This virus encodes human CD40L, a transmembrane type II protein expressed on CD4+ T cells of which the receptor is mainly expressed on antigen-presenting cells such as dendritic cells (DCs). Therefore, this OAd can activate DCs, which can then result in increased tumor-specific T cell responses. HAdV-B3 also played an important role when selecting OAds using directed evolution, representing an innovative method to generate novel OAds such as ColoAd1. This OAd was selected by co-infecting cells with several wild-type Ads and subsequent passaging allowing recombination events [54]. ColoAd1 mainly contains HAdV-B3 sequences, but in the E2B region, there are frequent substitutions with HAdV-B11 sequences (base pairs 6081 to 9322). Moreover, ColoAd1 is characterized by a nearly complete (2444 bp) E3 region deletion and a smaller (25 bp) second deletion that maps to a putative E4orf4 region. The same directed evolution strategy was applied in another study but here for selection ovarian cancer cells (SKOV3 cells) were seeded onto a Matrigel in a three-dimensional manner. Selection using Ads from species B-F of generated pools resulted in OvAd1 and OvAd2 [55], containing sequences of ColAd1 and HAdV-B3. In 2021, the first OAd was introduced solely based on HAdV-B35. This CAR-independent OAd Ad35-hTERT-E1A [56] was shown to be superior to a HAdV-C5-based OAd in various cancer cell lines. Another species B OAd is based on HAdV-B11 (RCAd11pADP), which was armed by expressing the Ad death protein in the E1 region [57]. This OAd showed enhanced lysis of metastatic prostate cells in vitro and in vivo.

The first non-HAdV-C5 tumor-specific species C viruses were developed in 2022. In this study, the Δ24 deletion was introduced in the E1 region of HAdV-C1, 2, 6, and in addition, these viruses were armed with the RNA interference inhibitor protein P19 derived from a tomato bushy stunt virus [58]. These OAds showed robust tumor cell lysis in vitro and enhanced tumor cell killing in vivo. Another OAd using HAdV-C6 as backbone is Ad6-hT-GM-CSF in which the E1A gene is expressed under the control of the hTERT promoter, and in addition, the virus was armed with GM-CSF, which was inserted into the E3 region (E36.7K and gp19K) [59]. This OAd displays robust oncolytic and immunostimulatory effects in a Syrian hamster model of cholangiocarcinoma.

Although species D represents the largest group of human Ads, OAds based on this species are scarce. In the year 2020, HAdV-D26, 28, 45, 48 were generated encoding the sodium iodide symporter protein (hNIS) inserted into E3 region [60]. These viruses were analyzed in a broad spectrum of cancer cell lines, resulting in robust replication. Another study investigated HAdV-D10 containing peptide A20 that selectively binds αvβ6 integrin [61]. Described species D viruses in these two studies display characteristics of a promising virotherapy, combining low seroprevalence and robust cancer cell lysis. However, tumor-specific replication remains to be solved by for instance inserting a tumor-specific promoter driving expression of the E1 region. Note that until today species A-, E-, F-, and G-derived Ads were not explored as OAds. However, with novel virus genome engineering technologies and the possibility to genetically modify basically any Ad type, it can be speculated that there will be more OAds available also from these species in the near future.

Numerous clinical trials were performed with OAds solely based on HAdV-C5 as backbone. The world’s first OAd, Oncorine (H101), was approved in 2005 in China and is based on HAdV-C5. Oncorine was investigated in a phase II clinical trial and showed potential in multicenter studies combined with chemotherapy especially for squamous cell carcinoma in head and neck and in nasopharyngeal cancer 7, 8. However, until today, there is a limited number of clinical trials using chimeric Ads or OAds based on Ads derived from a different Ad species. The first non-HAdV-C5 OAd (Ad3-hTERT-E1A) was used in a phase I clinical trial. Here, the OAd was intravenously injected in patients suffering from non–small cell lung cancer (NSCLC), breast cancer, sarcoma, pancreatic, prostate, bladder cancer, neuroblastoma, ovarian, and thyroid cancer [62].

Enadenontucirev (ColoAd1) was explored in several clinical studies since 2017 [63]. In a phase I clinical trial, the mechanism of action was investigated by intravenous delivery of ColoAd1 in patients with resectable carcinoma of the colon, patients with NSCLC, urothelial cell cancer, and renal cancer. Here, 17 patients obtained a single-dose treatment intratumoral with ColoAd1 at the first day of the study. After resection, an additional dose of ColoAd1 was given intravenously. The study provided feasibility of this approach, including intravenous delivery of ColoAd1. OAds NG-641 and NG-350A are based on ColoAd1 and were armed either with a protein-directed bi-specific T-cell activator antibody (FAP-Tac), CXCL9, CXCL10, IFNα, or an anti-CD40 antibody, respectively. NG-350A was injected into patients suffering advanced epithelial tumors and evidence for OAd delivery to the tumor and viral replication was provided [64].

In conclusion, the development of OAds is an exciting field and for the established vectors mainly based on HAdV-C5 and HAdV-B3 backbones high potential for tumor patients was demonstrated. However, although HAdV-C5 has its limitations, there are also barriers for using non-HAdV-C5 OAds, especially when planning to explore Ad types other than species C. Species C Ads, in particular HAdV-C2 and HAdV-C5, were characterized in detail in basic virology studies, including their replication cycle and virus–host interactions. Also for translational studies, mainly HAdV-C5-based vectors were explored. Therefore, there is in-depth information available for these species C-type HAdVs. However, basic knowledge about the majority of other HAdV types from other species regarding replication cycle and virus–host interactions is often scarce. As a consequence generation, production and upscaling of OAds derived from other HAdV species can be challenging. For instance, when considering a clinical trial establishment of a producer cell lines achieving high titers and pure vector preparations may be difficult. Furthermore, effectiveness of tumor cell killing can be a limitation and potential replication efficiencies need to be enhanced by arming these OAds with cell-killing enhancing factors.

The perspective to take advantage of the complete natural diversity of all 116 human Ads that were identified until today will open the path to further increase the effectiveness of OAds. We believe that with the advances in cloning techniques and DNA synthesis technologies generation of novel OAds based on any Ad type can be achieved. Numerous effector transgenes have been and are currently being tested to increase oncolytic potential of OAds as well as to trigger or increase antitumoral immune responses against the respective tumors.