WHAT IS ALREADY KNOWN ON THIS TOPIC

Oncolytic viruses have shown promise in enhancing immune responses and eliciting antitumor activity in various solid tumors.

The efficacy of oncolytic viruses, either as a single agent or in combination with immune checkpoint inhibitors, remains underexplored in soft-tissue sarcomas, a cancer type that often exhibits resistance to immune therapy.

WHAT THIS STUDY ADDS

The oncolytic herpes simplex virus-2 virus OH2 injection has demonstrated a favorable safety profile and significant antitumor activity in patients with advanced or metastatic sarcoma, particularly when used in combination with HX008 (a PD-1 inhibitor).

Post-treatment tumor specimens revealed immune modulation within the tumor microenvironment.

HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICYIntroduction

Soft-tissue sarcomas (STSs) are rare cancers of mesenchymal origin, accounting for 1% of all solid tumors.1 Conventional treatment for advanced STS is dominated by cytotoxic chemotherapy regimens. Over the last 15 years, immune checkpoint inhibitors (ICIs) have emerged as a revolutionary oncologic therapy. The SARC028 and ALLIANCE (A091401) trials demonstrated antitumor activity of ICIs2 3 in some subtypes of sarcoma, such as undifferentiated pleomorphic sarcoma (UPS) and alveolar soft part sarcoma (ASPS). However, primary and acquired resistance to ICI often reduces clinical benefits in STS and other cancer indications. Therefore, there is an unmet need to effectively combine ICIs with other treatments to improve efficacy and survival in sarcoma. One such class of agents could be oncolytic viruses (OVs). Patient treatment with OVs triggers antiviral immune responses, releasing cytokines and chemokines, which attract immune cells in the tumor microenvironment (TME).4 OVs replicate selectively in tumor cells leading to tumor cell killing that generates tumor-associated antigens and tumor neoantigens, stimulating adaptive immunity5 and potentially transforming the tumor from “(ICI-)cold” to “(ICI-)hot”.6 The efficacy of OVs in sarcoma has been investigated through various administration methods, including intratumoral or intravenous approaches, and in combination with ICIs or radiotherapy.7–9 Talimogene laherparepvec (T-VEC) was the first OV employed in sarcoma patients, administered intratumorally in combination with intravenous pembrolizumab. In a phase 2 clinical trial involving 20 patients, T-VEC achieved an overall response rate (ORR) of 35%, with partial responses (PR) observed in cutaneous angiosarcoma (AS), UPS, myxoid fibrous sarcoma, epithelioid sarcoma and unclassified sarcoma.7 Another study evaluating the efficacy of intratumoral JX-594 combined with avelumab and low-dose cyclophosphamide reported one response in a patient with AS.10

OH2 is a genetically engineered OV derived from the wild-type herpes simplex virus (HSV)-2 strain HG52, constructed with the addition of human granulocyte-macrophage colony-stimulating factor (GM-CSF) encoding gene and the deletion of ICP34.5 neurovirulence gene. HX008 is an anti-programmed cell death protein 1 (anti-PD-1) monoclonal antibody.11

OH2 was previously tested alone or in combination with HX008 in patients with advanced solid tumors in a multicenter, phase 1/2 clinical trial. The combination demonstrated durable antitumor activity in metastatic esophageal and rectal cancer.12 Here, we report the results of a phase 1/2 study in patients with advanced sarcomas to investigate the safety and antitumor activity of OH2 as a single agent or in combination with HX008. We also report alterations of TME and prognostic immune factors that we identified using tumor samples obtained before and after OV injection.

MethodsPatients

This was a phase 1/2, open-label, non-randomized, multicenter trial, conducted in patients diagnosed with locally advanced or metastatic STS, with documented disease progression within 6 months prior to entry into the study. Key inclusion criteria were age ≥18 years; having 1 or more injectable lesions (longest diameter ≥5 mm; measurable disease per RECIST V.1.1 outside any previously irradiated field; adequate hematological, renal, metabolic and hepatic functions; having received 1 or more lines of standard systemic therapy; having an Eastern Cooperation Oncology Group performance status of 0 or 1. Key exclusion criteria were symptomatic autoimmune disease; a documented allergic history to HSV; clinical manifestations of active infection or unexplained fever exceeding 38.5°C prior to study commencement; having known active central nervous system metastasis and tumor volumes insufficient to accommodate the minimum amount of OV injection. The authors vouch for the accuracy and completeness of the data.

Procedures

The phase 1/2 study consisted of a single-agent OH2 cohort and a OH2/HX008 combination cohort. Patients were assigned to receive either single-agent OH2 or OH2 plus HX008. OH2 was administered intratumorally at 3 dose levels (106, 107 and 108 CCID50/ml) every 2 weeks, whereas in the combination cohort HX008 was administered intravenously at a fixed dose of 200 mg every 3 weeks, in accordance with the approved label of HX008 in China. In the first cycle, the two agents were administered on the same day (day 1), then administered according to their own schedules (OH2 every 2 weeks, HX008 every 3 weeks). Dose escalation of OH2 followed the traditional “3+3” design, which was described in a previous study.12 If two or more dose-limiting toxicities (DLTs) were documented at one dose level, the maximum tolerable dose (MTD) would be defined as the previous lower dose.

Intratumoral OH2 injections were performed either directly for cutaneous or subcutaneous lesions, or with ultrasound guidance for deep-located nodes or organ metastases. The volume injected into each lesion was based on the longest diameter of the tumor (≤1.5 cm, up to 1 mL; >1.5 to ≤2.5 cm, up to 2 mL; >2.5 to ≤5.0 cm, up to 4 mL; >5.0 cm, up to 8 mL). Injection of multiple tumors was allowed. There were no limits on the number of lesions that could be injected per patient, but the maximum volume that could be injected during each visit was set to 8 mL.(online supplemental file 1)

Response evaluation and safety

Tumor assessments were conducted at screening, every 6 weeks during the first 6 months after treatment initiation, every 8 weeks for the next 6 months, and then every 12 weeks for the next 12 months. Radiographic assessments were performed using CT or MRI. Measurement of cutaneous or subcutaneous lesions was conducted with calipers. Evaluation of response was performed by the investigators using both the RECIST V.1.1 and the immune-RECIST (iRECIST) criteria. Toxicities were evaluated continuously per Common Terminology Criteria for Adverse Events V.5.0. DLT was defined as any of the following events within the first 3 weeks of therapy that was deemed treatment related by the investigator(s): grade 4 neutropenia lasting over 5 days or febrile neutropenia; grade 3 thrombocytopenia with bleeding; any other grade 4–5 hematological adverse events (AEs); grade 3 rash over 3 days; grade 3 nausea, vomiting and diarrhea lasting over 3 days despite appropriate supportive care; grade 3 alanine aminotransferase or aspartate aminotransferase level increase lasting over 7 days; grade 2 proteinuria lasting over 7 days; any other grade 3–5 non-hematological AEs.

Outcomes

The primary endpoint was safety and tolerability for the phase 1 portion and ORR determined by RECIST (V.1.1) and iRECIST criteria for the phase 2 of the trial. Secondary endpoints included duration of response (DOR), progression-free survival (PFS), and overall survival (OS).

Immunohistochemical analysis

To evaluate the effects of OH2 on the TME of sarcoma, immunohistochemical staining was performed using anti-CD3 (GR107, Shanghai GeneTech, Shanghai, China), CD4 (EP132, Zhongshan Golden Bridge Biotechnology, Beijing, China), CD8 (SP16, Zhongshan Golden Bridge Biotechnology, Beijing, China), CD20 (L26, Agilent Technologies, Santa Clara, California, USA), FoxP3 (ab20034, Abcam, Cambridge, UK), CD56 (123C3.D5, Zhongshan Golden Bridge Biotechnology, Beijing, China), and PD1 (UMAB199, Aorui Dongyuan Biotechnology, Wuxi, China) primary antibodies. PD-L1 expression was assessed using Combination Positive Scores (CPS) after immunohistochemistry staining on 4 µm thick FFPE sections using PD-L1 22C3 pharmDx (Dako, Carpinteria, California, USA). To evaluate the expression of the virus genome following OH2 treatment in the sarcoma tissue, immunohistochemical staining of GM-CSF (1:400, 17 762–1-AP, Proteintech, China) expression was performed in pretreatment and post-treatment paired samples. The staining was carried out on the Leica BOND III automated immunostainer (Leica Biosystems, Newcastle, UK) using the Bond Polymer Refine Detection Kit (Leica Biosystems, #DS9800). All slides were scanned with a Pannoramic P250 FLASH slide scanner (3DHistech, Budapest, Hungary), and the number of positive cells was counted automatically using the QuPath software V.0.5.0 (Edinburgh, UKm). Furthermore, the necrosis was morphometrically analyzed on HE-stained tumor slides, and the necrosis rate per every slide was quantified as necrotic area divided by total area. The average necrotic rate of each case was calculated by arithmetical averaging.

Statistics analysis

The efficacy analysis included patients who received study therapy and underwent evaluation for response. Safety analysis included patients who received at least 1 dose of treatment. Categorical data analysis and summary statistics were used to report AEs. The Kaplan-Meier method was used to estimate time-to-event variables. PFS was defined as the time from the start of treatment to the time of progression or death (from any cause). OS was defined as the time from the start of treatment to death (from any cause) or the last patient contact. For other variables such as demographic data, continuous variables are represented as median and range, and categorical variables are represented as percentage. Pretreatment and post-treatment comparisons of density of immune cells were calculated using a two-tailed Mann-Whitney test (p <0.05).

Results

Between October 20, 2020 and December 30,2023, 26 patients were enrolled in this study. Seven patients were treated with single-agent OH2 and 19 with OH2 in combination with HX008. The baseline demographic characteristics of the patients are summarized in table 1. Two (28.6%) and six patients (31.6%) had previously received ICIs in the single OH2 and combination groups, respectively. The median follow-up time for survival analysis was 11.9 months. 25 patients discontinued treatment: 20 due to progressive disease, 3 withdrew consent, and 2 because of AEs.

Table 1

Baseline characteristics

All patients were included in the safety analysis. AEs were common but well tolerated in both groups. No DLTs were observed and the MTD was not reached during the dose escalation. The most frequent treatment-related AEs (TRAEs) occurring at any grade in 20% or more of patients were fever (57.1%), anemia (28.6%) and decreased white cell count in the single-agent group; fever (26.3%), γ-GGT elevation (26.3%) and weight loss (21.1%) in the combination group. Grade 3 or 4 AEs were reported in two patients (28.6%) in the single-agent group (anemia) and two patients (10.5%) in the combination group (including hypertriglyceridemia, nausea, neck pain and anorexia). There were no serious TRAEs or treatment-related death throughout the study. The TRAEs documented in the study are summarized in online supplemental tables S1 and S2.

Seven patients from OH2 single-agent group and 18 patients from OH2 plus HX008 group were evaluated for response (one from the combination group withdrew consent before the first evaluation). Responses of injected lesions were observed in four patients receiving OH2 plus HX008 (2 complete responses (CRs) and two PRs), while no response was noted in patients treated with single-agent OH2 (figure 1). 12 patients had measurable distant disease beyond the injectable tumor sites, with responses of non-injected lesions observed in 1 patient receiving the combination regimen (online supplemental figure S1). According to RECIST V.1.1, considering both injected and non-injected lesions, the ORR was 0% and 16.7% in the OH2 single agent and OH2 plus HX008 group, respectively.

Figure 1

Response to the treatment. (A) Waterfall plot showing the best change of injected lesions from baseline. The square indicates one patient having a shrinking injected lesion who nevertheless stopped treatment due to a growing non-injected lesion. The arrows indicate patients having complete or partial response of injected lesion but who also stopped treatment due to the occurrence of a new lesion. The star indicates one patient having a complete response of all lesions to date. (B) Spider plot showing changes of injected lesions from baseline.

The median PFS for the single agent OH2 and OH2/HX008 combination was 1.41 (95% CI 1.15 to 2.73) and 1.45 (95% CI 1.38 to 5.36) months, respectively. At the data cut-off, 16 patients had died (6 in the single-agent group and 10 in the combination group). The median OS for the single-agent and combination group was 4.5 and 18.0 months, respectively (figure 2).

Figure 2

Kaplan-Meier plot of progression-free survival (PFS) (A) and overall survival (OS) (B).

Responsive tumor subtypes included one fibrosarcoma (FS), one liposarcoma and two AS (online supplemental table S3). The patient with liposarcoma experienced tumor shrinkage of injected lesions but discontinued treatment due to a significant growth of non-injected lesions. The other three patients had locally advanced disease, with DOR of 3.9 months, 5.5 months and 6.5 months, respectively. Notably, one responding AS patient had failed paclitaxel plus ICI before entering the study, implying a synergetic effect of OH2 with HX008. This patient remained in treatment at the data cut-off. In the OH2 plus HX008 group, one patient with ASPS had a 12% shrinkage of the injected lesion in the liver, with stable disease of non-injected lesions in the lung.

We employed immunohistochemical methods to explore the status of tumor-infiltrating lymphocytes (TILs) in nine samples from six patients in the OH2/HX008 group. Three of these patients had paired pretreatment and post-treatment samples. The post-treatment tumor sample from 12S012 was excluded from subsequent statistical analyses due to extensive necrosis and minimal immune cell infiltration. Consequently, a total of eight samples were analyzed, comprising four pretreatment and four post-treatment samples. The results indicated that post-treatment with OH2 led to increased densities of CD3, CD4, CD8, and PD-1 cells, while the density of FoxP3 cells decreased. However, only the difference in CD8 density reached statistical significance (p=0.043) (figure 3). We further examined the changes in TILs in two patients (12S007 and 12S001) who had paired samples. Post-treatment, the densities of CD4−, CD8−, and PD1-positive cells increased, whereas the density of FoxP3-positive cells decreased (online supplemental figure S2). Morphological observation of H&E-stained sections revealed varying degrees of necrosis in all five post-treatment samples, with necrosis ratios ranging from 5% to 95% (online supplemental figure S2). The expression of GM-CSF was analyzed in three paired samples. The results demonstrated that GM-CSF expression intensity increased to varying extents across all samples following treatment, with 12S007 showing a particularly marked increase in staining intensity (online supplemental figure S3). Notably, this patient achieved a PR during treatment. PD-L1 expression was evaluated using CPS following immunohistochemical staining. Only one sample demonstrated high PD-L1 expression post-treatment, with a CPS of 20. However, the absence of a paired pretreatment sample precludes any conclusions regarding whether this elevated expression was a result of the treatment. The remaining cases exhibited CPS scores either below 1 or at 0, indicating a limited PD-L1 response to treatment across the cohort (online supplemental table S4).

Figure 3

Comparison the density of immune cells pre-OH2 and post-OH2 therapy (horizontal bar: median). (A) CD3; (B) CD4; (C) CD8; (D) PD1; (E) FoxP3. *Significant difference (p<0.05).

Discussion

Oncolytic virotherapy has been validated in multiple malignant tumors, including glioblastoma,13 14 melanoma15 and other solid tumors.16 For sarcoma, the therapeutic effect of genetically engineered viruses, such as HSV-1 and vaccinia virus, has been demonstrated in recent clinical trials.7 10

In the present phase 1/2 trial, we reported that the oncolytic HSV-2 virus OH2 was well-tolerated, and we provided more convincing evidence of OH2/HX008 efficacy in sarcoma patients.

The most reported TRAE in our trial was transient fever, which was managed using antipyretics. The incidence of grade 3 or 4 TRAEs was low (15.4%), with no serious TRAE reported. This safety profile is consistent with previous experience with OH2 in patients with solid tumors,12 and comparable with the safety profile of most other OVs administered intratumorally, including HSV-1-based T-VEC,7 adenovirus based DNX-2401,7 and vaccina virus.17

Tumor responses were documented in 4 out of 25 patients per RECIST V.1.1 treated with the combination OH2/HX008. In a recent study, where intratumoral T-VEC and intravenous pembrolizumab were coadministered in patients with advanced STS,7 the best ORR was significantly higher (30%) than the data from our study and historical ORR of PD-1 monotherapy (ranging from 5% to 18%).2 3 18 Notably, responses in our study were predominantly observed in several tumor subtypes, including AS and FS. In another study, where intratumoral JX-594, an engineered poxvirus, was coadministered with avelumab and low-dose cyclophosphamide, objective response was documented in only one AS patient.10 Compared with the TVEC study,7 the authors of JX-594 study attributed the result to variations in the TME at baseline, including differences in the abundance of TILs. Previous research showed that TILs were enriched in AS, FS and UPS.19–23 When considering these tumor subtypes separately in our study, the ORR was comparable with T-VEC (3/6 vs 7/15), indicating that selecting specific tumor types might be necessary to demonstrate OH2 efficacy in future studies.

The trial was designed to enroll patients with locally advanced disease, which is usually unresectable without severe morbidity. Among the 11 patients with locally advanced disease, 3 patients (27.3%) achieved PR or CR. Nowadays, neoadjuvant chemotherapy is recommended for patients with localized, large and high-grade STS. The response rate of neoadjuvant regimens ranges from 10% to 30%.24–27 Consistent with this, neoadjuvant oncolytic virotherapy has been explored in several cancers including melanoma and breast cancer, leading to tumor eradication and prolonged disease-free survival.28 29 Moreover, preoperative intra-tumoral T-VEC injection with concurrent radiation in sarcoma patients led to enhanced immune responses with a satisfactory safety profile.8

Besides direct oncolysis of growing tumors after intratumoral OV injections, OVs cause the release of tumor-associated antigens in the TME, which may stimulate systemic immune responses leading to elimination of distant tumor cells.30 Such an “abscopal effect” response was reported in a randomized phase 2 trial comparing T-VEC plus ipilimumab with ipilimumab monotherapy in metastatic melanoma patients.31 Similarly, in the sarcoma trial combining JX-594 with avelumab and low-dose cyclophosphamide a significant elevation of CD8+TILs in non-injected lesions was documented after 2 cycles of treatment,10 consistent with an abscopal effect of this vaccinia OV as well. In contrast, in our case, regression of distant non-injected lesions was documented only in one patient on the combination regimen, yet was paradoxically associated with an increase in injected lesion volume, which contrasts the known OV abscopal effects. The basis of this observation is unclear at present, but the putative abscopal effect (or lack thereof) of OH2 in sarcoma cannot be confirmed in the absence of a control group, which would call for a randomized trial comparing OH2 plus ICI versus ICI monotherapy.

Our TILs analysis in sarcoma pre-OH2 and post-OH2 treatment demonstrate significant activation of the cellular immune response by OH2/HX008. In line with other studies, we observed a marked increase in CD8 T-cell density, also known as cytotoxic T-cells, a fraction that is capable of mediating tumor regression through direct tumor cell lysis involving perforin and granzyme release in the TME.8 10 Although the increases in CD3, CD4 T-cell and PD-1 cell densities did not reach statistical significance, their elevated presence post-treatment suggests an overall shift toward a more active immune microenvironment. Conversely, the decrease in FoxP3 cells, indicative of regulatory T-cells (Tregs), is consistent with a favorable shift in the immune balance toward lower immune suppression mediated by Tregs and higher antitumor immune response.32 The extensive necrosis observed in post-treatment samples, ranging from 5% to 95%, indicates varying degrees of response to OH2/HX008 therapy including direct lysis and remodeling the TME to support a robust immune-mediated tumor suppression.

Limitations

Although this trial achieved positive results, there are some limitations. First, adequate tumor materials were not retrieved from all enrolled patients, which limits correlative analyses. Second, the possibility of delayed responses despite an apparent initial patient progression (“pseudoprogression”) should be considered in STS as well.8 Pseudoprogression was observed in 50% of patients in a phase II melanoma trial.33 34 The necrosis rate after surgery implies these patients may benefit from further OH2 treatment. Third, without a control group, it is hard to discern the benefit of the combination therapy from either drug alone. However, one responsive patient in our study was refractory to previous ICI, and another patient with ASPS had a shrinkage of the injected lesion but no responses in non-injected lesions. These results suggest OH2 may synergize with ICI to elicit clinical benefit.

Conclusion

Intratumoral injection of OV OH2 was safe in patients with sarcoma. Encouraging antitumor activities were observed in patients with AS, and FS for the OH2/HX008 combination. The effectiveness and low toxicity favor further investigation of OH2 and HX008 in neoadjuvant regimen in selective STS subtypes.

Data availability statement

Data are available on reasonable request. Data are available on reasonable request, please contact liujiayong_doc@163.com.

Ethics statementsPatient consent for publication

Not applicable.

Ethics approval

This study involves human participants and was approved by the Peking University Cancer Hospital & Institute ethics committee (Approval number: 2021YW157). Participants gave informed consent to participate in the study before taking part.

Acknowledgments

All authors have read the authorship policy for the journal. We are grateful to all the patients who generously volunteered to participate in this study. We thank the research nurses and study coordinators for their efforts on behalf of the patients. We appreciate the assistance provided by Binhui Biopharmaceutical for this clinical trial.