Article Text

Original research
Peritumoral administration of immunomodulatory antibodies as a triple combination suppresses skin tumor growth without systemic toxicity
  1. Quentin G Wright,
  2. Debottam Sinha,
  3. James W Wells,
  4. Ian H Frazer,
  5. Jazmina L Gonzalez Cruz and
  6. Graham Robert Leggatt
  1. Frazer Institute, The University of Queensland, Brisbane, Queensland, Australia
  1. Correspondence to Dr Graham Robert Leggatt; g.leggatt{at}


Background Skin cancers, particularly keratinocyte cancers, are the most commonly diagnosed tumors. Although surgery is often effective in early-stage disease, skin tumors are not always easily accessible, can reoccur and have the ability to metastasize. More recently, immunotherapies, including intravenously administered checkpoint inhibitors, have been shown to control some skin cancers, but with off-target toxicities when used in combination. Our study investigated whether peritumoral administration of an antibody combination targeting PD-1, 4-1BB (CD137) and VISTA might control skin tumors and lead to circulating antitumor immunity without off-target toxicity.

Methods The efficacy of combination immunotherapy administered peritumorally or intravenously was tested using transplantable tumor models injected into mouse ears (primary tumors) or subcutaneously in flank skin (secondary tumors). Changes to the tumor microenvironment were tracked using flow cytometry while tumor-specific, CD8 T cells were identified through enzyme-linked immunospot (ELISPOT) assays. Off-target toxicity of the combination immunotherapy was assessed via serum alanine aminotransferase ELISA and histological analysis of liver sections.

Results The data showed that local administration of antibody therapy eliminated syngeneic murine tumors transplanted in the ear skin at a lower dose than required intravenously, and without measured hepatic toxicity. Tumor elimination was dependent on CD8 T cells and was associated with an increased percentage of CD8 T cells expressing granzyme B, KLRG1 and Eomes, and a decreased population of CD4 T cells including CD4+FoxP3+ cells in the treated tumor microenvironment. Importantly, untreated, distal tumors regressed following antibody treatment of a primary tumor, and immune memory prevented growth of subcutaneous flank tumors administered 50 days after regression of a primary tumor.

Conclusions Together, these data suggest that peritumoral immunotherapy for skin tumors offers advantages over conventional intravenous delivery, allowing antibody dose sparing, improved safety and inducing long-term systemic memory. Future clinical trials of immunotherapy for primary skin cancer should focus on peritumoral delivery of combinations of immune checkpoint antibodies.

  • Skin Neoplasms
  • Immunotherapy
  • Immunomodulation
  • Tumor Microenvironment

Data availability statement

Data are available on reasonable request. The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.

This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.


  • Intravenously administered antibodies against checkpoint molecules such as PD-1 and CTLA-4 are being used in the clinic to treat skin cancers. However, a subset of tumors can be resistant to this therapy and off target toxicities can be problematic particularly when combinations of antibodies are used. There is a clear clinical need for optimizing new combinations of checkpoint antibodies with improved efficacy that can be delivered in a safe manner for the treatment of skin cancer.


  • We have identified a triple combination of antibodies targeting inhibitory and stimulatory checkpoint molecules that can successfully eliminate three different cutaneous tumors after peritumoral administration. Local administration enabled the use of lower antibody doses, prevented off target toxicities in the liver and was associated with a systemic, memory immune response which protected against tumor challenge at distant skin sites.


  • Peritumoral administration of novel checkpoint antibody combinations could prove to be a potent therapeutic option in the future as standard clinical practice in the treatment of primary skin cancers.


Cutaneous cancers such as squamous cell carcinoma (SCC) and melanoma represent a spectrum of disease from localized in situ carcinoma of the skin through to widely dispersed, metastatic tumors.1 The global incidence of these cancers is increasing with non-melanoma skin cancer (NMSC) being the most common malignancy in individuals with Fitzpatrick skin types I–III.2 3 While treatment of early-stage cancer with surgery is often effective, the 5-year recurrence rates for cutaneous SCCs are 4%–8% and highly susceptible patients often have multiple lesions, some of which may not be easily accessed.4 5 Other treatment options such as topical chemotherapeutics (imiquimod) and radiotherapy have varying effectiveness and undesirable side effects and may not confer lasting protection.6 7 A role for the immune system in controlling skin tumors is implied by the higher incidence of NMSC in immunocompromised individuals, particularly organ transplant recipients where the incidence of cutaneous SCC is 65–250 times higher than the general population.8 9 These cancers also have high mutational burdens (induced by UV light) suggesting the possibility of unique cancer antigens which can be targeted by the adaptive immune system.10 Recently, tools for manipulating the immune response have expanded with the identification of surface checkpoint molecules which signal to inhibit (eg, PD-1) or costimulate (eg, 4-1BB) immune cells. Food and drug administration approvals for the use of checkpoint inhibitor antibodies, which block PD-1 and CTLA-4 signaling, in cutaneous cancers have ushered in a new era of therapy, pivoting the focus of treatment from targeting tumors or cancer cells, to harnessing the immune system, as means to eradicate tumors.11–13 While these antibodies have shown promising efficacy in skin cancer, treatment resistance to monotherapies can occur and treatment is generally applied to advanced or metastatic cancer due to initial failure of “standard-of-care” treatments or concerns related to systemic toxicities when using antibodies against molecules such as CTLA-4 or 4-1BB. Autoimmune activation is particularly problematic when combinations of antibodies change the signaling in multiple non-redundant signaling pathways.14 Intravenous delivery of antibodies against both 4-1BB and PD-1 exacerbated the liver toxicity seen with anti-4-1BB antibody alone in a mouse model, suggesting that safer delivery methods for combinations of checkpoint antibodies are needed.15

To enable the use of combination antibody therapy targeting checkpoint molecules in skin cancer, our study has investigated the peritumoral injection of a novel combination of antibodies that block the inhibitory signals from PD-1 and VISTA while simultaneously promoting activation signals through 4-1BB. These targets were chosen on the basis of (1) their expression within human, cutaneous SCC (2) a role for VISTA in resistance to PD-1 therapy and (3) their non-redundant signaling pathways in immune cell activation.16 Overall, our data demonstrate that peritumoral triple therapy is highly effective in regressing different types of cutaneous tumors with the establishment of systemic immune memory and protection against distant, simultaneous tumors. Importantly, therapy can be achieved at lower antibody doses than required intravenously and with no evidence of systemic toxicity. Together, this suggests a new paradigm for the clinical treatment of early stage, primary skin cancer with multiple antibodies against immune checkpoint molecules.



C57BL/6 or RAG1KO (B6.128S7-Rag1tm1.Mom/J) adult (6–8 weeks old) mice were sourced from the Animal Resource Centre (Perth, Australia) or the University of Queensland Biological Research Facility (UQBRF) at the Translational Research Institute (Brisbane, Australia). All mice were bred and housed under specific pathogen free (SPF) conditions at the UQBRF. Mice were regularly monitored for signs of distress and any adverse events were reported immediately to the ethics committee.

Tumor cell lines and in vivo establishment of ear and subcutaneous tumors

TC-1 lung epithelial, B16F10 melanoma and 8101-PRO UV-induced fibrosarcoma cell lines have been previously described.17–19 These cells were maintained for 3–4 passages in complete RPMI media containing FCS prior to injection. For establishment of tumors, TC-1 cells were trypsinized and washed to prepare a single cell suspension (>85% viability). TC-1 cells resuspended in PBS were mixed with an equal volume of Matrigel (Phenol red free, Corning, cat#356237) before injection into the ventral ear pinnae of C57BL/6 mice at 5×105 cells in a 20 µL total volume. For B16F10 and 8101-PRO tumors, 1×106 cells were injected into the ear in Matrigel. Tumor volumes in the ear were monitored daily and measured using digital calipers. For tumors grown in back skin, 5×105 tumor cells were placed in a 200 µL volume of PBS and injected subcutaneously into the shaved lower back. Subcutaneous placement of the tumor was confirmed with histology. Tumor volumes were calculated using the following equation as previously described: volume (mm3)=(length×width2)/2.20 For histological analysis, mice were euthanized and excised, tumor injected ears (or flank tumors) and livers were placed in 4% paraformaldehyde before processing into H&E sections by the histology core facility of the Translational Research Institute (Brisbane, Australia).

Antibodies, in vivo therapy and flow cytometry

For tumor therapy experiments, unlabelled antibodies against PD-1 (Clone RMP1-14; Rat IgG2A)), 4-1BB (Clone 3H3; Rat IgG2A), VISTA (Clone 13F3; Armenian Hamster IgG), Rat IgG2A isotype control (Clone 2A3), Armenian hamster isotype IgG, CD8β (Clone 53-5.8; Rat IgG1) and Rat IgG1 isotype control (Clone HRPN) were all purchased from BioXcell (New Hampshire, USA). For multiparameter flow cytometry, fluorescentlyconjugated antibodies targeting CD45 (Clone 30-F11; BUV563), CD49b (Clone HM-α2; BV786), TCR-b (Clone H57-597; BUV737), CD4 (Clone GK1.5; BUV 496), CD8α (Clone 53-6.7, BUV805) and live/dead stain (BUV450/50) were all purchased from BD (New Jersey, USA). Antibodies against FoxP3 (Clone FJK-16s; PE/Cy5), KLRG-1 (Clone 2F1; FITC), and Eomes (Clone Dan11mag; PE/Cy7) were purchased from Thermofisher (Massachusetts, USA) and anti-granzyme B antibody (Clone QA16A02; PE/Dazzle 594) was purchased from Biolegend (San Diego, USA).

For flow cytometry staining of cells from ear tumors, ears were excised from euthanized mice and ventral/dorsal surfaces were separated mechanically, minced and then placed in 100 µg/mL Liberase (Roche, Basel, Switzerland) and 50 µg/mL DNase I (Sigma, St. Louis, USA) in DMEM solution for 40 mins at 37°C. Finally, the cell suspension was passed through a 70 µm cell strainer and incubated for 20 min with the live/dead stain (BD) and 1 µL of TruStain FcX (Biolegend) to block Fc binding. Cells were then incubated with antibodies targeting surface molecules for 20 mins in 100 µL of PBS before being washed and analyzed in the flow cytometer. For intracellular FoxP3 staining, cells were fixed and permeabilized using the FoxP3 Transcription factor staining kit as per manufacturer’s instructions (Thermofisher) and incubated overnight with anti-FoxP3 antibody. Multicolour cell fluorescence and compensation beads (Ultracomp beads; Thermofischer) were acquired on a BD LSR Fortessa X-20 (BD) and the data analyzed using Flowjo V.10 (BD).

For in vivo antibody therapy experiments, antibodies were injected on three separate occasions spaced 3 days apart beginning on day 3 when a small tumor mass was palpable (days 3, 6, 9 post-tumor inoculation). For ear injections, antibodies (10 µg per antibody; mixed together for combination treatment) were injected beside the tumor mass in a volume of 20 µL PBS while intravenous injection (10 µg or 100 µg per antibody) was administered in the tail vein in a total volume of 200 µL. In abscopal experiments, tumor was injected into both ears of an individual mouse but antibody therapy was applied to the right ear only. In experiments requiring CD8 T cell depletion, 250 µg of antibody targeting CD8β (Lyt 3.2, clone 53-5.8, BioXcell) was administered intraperitoneally 3 days prior to TC-1 tumor inoculation and again 4 days after tumor inoculation. CD8 T cell depletion in the blood (>95% depletion at the time of tumor inoculation) was confirmed by flow cytometry.


Harvested spleens were processed into single cell suspensions. Millipore Multiscreen-HA 96-well plates precoated with 4 µg/mL anti-IFN-γ mAb (eBioscience #14-7313-85), were washed and blocked with sterile RPMI 1640 full media (RPMI 1640/penicillin/streptomycin/glutamine/sodium pyruvate+50 µM 2-mercaptoethanol+10% fetal bovine serum), for 2 hours. Blocking solution was then removed and 2.5×105 splenocytes were added to triplicate wells for peptide and no peptide control conditions, in a volume of 50 µL. A 50 µL of RPMI full media was added to no peptide controls. A 50 µL of the immunodominant, HPV16E7 peptide (Auspep; RAHYNIVTF)), was added to peptide stimulated wells at a concentration of 20 µg/mL. PMA/ionomycin (50 ng/mL and 2 µg/mL, respectively) served as an assay positive control and was added in triplicate at 50 µL/ well. Cells were incubated for 24 hours at 37°C, 5% CO2. The contents of the plate were then aspirated and washed 3X with PBS/Tween (PBS-T). Biotinylated IFN-γ detection antibody (eBioscience #13-7312-85) was added to the plates at a concentration of 1 µg/mL in 2% FBS PBS-T. The plate was incubated for 4 hours at room temperature and later washed 3X with 2% FBS PBS-T. Avidin-HRP (Sigma cat # A- 3151, 1 mg/mL stock) was applied to the plate at a concentration of 2.5 µg/mL and incubated in the dark for 60 min at room temperature. The plate was washed 3X with PBS-T and 3X with PBS before adding DAB substrate (Sigma cat # DO426). Plates were developed for approximately 90–120 s. Color development was stopped by removing the plate backing and washing thoroughly in water. The plate was then allowed to dry prior to being read on a ELISPOT counter.

Serum alanine transaminase assay

Changes in serum alanine transaminase (ALT) levels were assessed by ALT protein ELISA (Abcam, ab282882), according to manufacturer’s instructions. Mean baseline mouse serum ALT levels were defined as 67.60 ng/mL.

Study design and statistical analysis

Sample sizes for tumor experiments (n=4–13) were chosen to minimize mouse usage while providing sufficient statistical power given the large effect sizes observed with treatment. No mice were excluded from the data analysis, the studies were not blinded and mice were randomized to study groups prior to tumor injection. Confounders such as order of treatments and measurements were not controlled. All statistical analyses were performed using Graphpad prism and summarized data are presented as mean±SD. Details of specific statistical tests are included in the figure legends. In experiments to compare the effect of monotherapy and combination therapies on mouse survival, a Kaplan-Meier survivability curve was generated. A Mantel-Cox test was applied to determine which treatments imparted significant survivability to test subjects compared with isotype control. P values less than 0.05 were considered significant and were numerically annotated in figures. Where asterisks were used, p<0.05 (*) p<0.01 (**), p<0.001 (***), and p<0.0001 (****).


Peritumoral injection of a combination of 4-1BB agonist and PD-1/VISTA checkpoint inhibitors induces superior antitumor responses when compared with monotherapy

To evaluate the efficacy of antibodies specific for 4-1BB, PD-1 and VISTA for clearance of a syngeneic tumor (TC-1) expressing a defined tumor antigen (HPV-16 E7 protein), each antibody was injected peritumorally around a small TC-1 tumor grown under the skin of the ear. Tumors in mice recipient of isotype control antibodies continued to grow until ethical limits were reached (figure 1A). Of nine mice recipient of antibodies against 4-1BB, seven became tumor free and remained so for greater than 50 days, whereas antibody against PD-1 eliminated tumor from one of nine mice, and delayed tumor growth for two mice, and antibodies against VISTA eliminated tumor from three of nine mice (figure 1A). Of the mice recipient of VISTA antibody whose ear tumors had regressed, two developed secondary tumors in tumor draining lymphatics within the neck (figure 1A). We next investigated the efficacy of peritumoral administration of a combination of antibodies specific for 4-1BB, PD-1 and VISTA. Of TC-1 tumor-bearing mice treated with this combination, all (8/8) became tumor free and remained so for a period of at least 60 days, whereas treatment with a combination of isotype control antibodies had no impact on tumor growth (figure 1B). Differences between anti-4-1BB monotherapy and triple antibody combination did not reach statistical significance. However, survival data suggested that combination treatment was superior to monotherapy (figure 1C). The combination of three antibodies was used in subsequent experiments.

Figure 1

A peritumorally administered combination of anti-4-1BB/PD-1/VISTA antibody therapy induces a superior antitumor response relative to monotherapy in an ear tumor model. (A, B) C57BL/6 J mice were injected with TC-1 tumor cells in the ear at day 0 followed by three injections of antibody monotherapy targeting the indicated molecule (A) or the triple combination of antibodies (B) at days 3, 6, 9 (indicated by dashed vertical lines). Isotype antibodies were used as controls. Tumor volumes for individual mice over time are plotted and the number of ear tumor free mice indicated. Mice whose tumors reached ethical volume limits (30–40 mm3) were terminated. (C) Monotherapy and combination therapy from A+B is summarized in a Kaplan-Meier survival curve. Data were pooled from two independent experiments and statistical significance was calculated using a log-rank (Mantel-Cox) test, *p<0.05, n.s.=not significant.

CD8 T cells are required for the antitumor efficacy of triple antibody treatment

CD8 T cells play a central role in the clearance of subcutaneous TC-1 tumors.21 To determine whether T and B cells played a part in clearing TC-1 tumor in our model, combination antibody therapy was administered to TC-1 ear tumor bearing immunocompetent C57BL/6 mice, and to T and B cell-deficient, RAG1−/− (RAG1KO) mice. As expected, in control immune competent mice, tumors grew progressively until ethical limits after isotype antibody treatment, whereas tumor was cleared after triple antibody treatment (figure 2A). In contrast, tumors in RAG1KO mice treated with triple antibody therapy demonstrated delayed growth, relative to isotype control treated animals, but continued to grow (figure 2A). To establish whether CD8 T cells contributed to tumor clearance, antibodies against the CD8β chain were used to deplete CD8 T cells from immunocompetent mice. This resulted in a >95% reduction in blood CD8 T cells (data not shown). CD8 T cell depleted animals given triple antibody therapy failed to clear TC-1 tumors (figure 2B), showing that tumor clearance in the ear induced by combination antibody therapy is CD8-dependent. In addition, we have shown that tumor antigen-specific, CD8 T cells are generated by day 12 after treatment of TC-1 tumors as evidenced by IFN-γ producing T cells in an ELISPOT assay after HPVE7 peptide stimulation (online supplemental figure 1).

Supplemental material

Supplemental material

Figure 2

Peritumoral antibody therapy is dependent on CD8+ T cells. (A) C57BL/6 mice and RAG1KO mice (lacking T and B cells) were injected with TC-1 cells followed on day 3 by three peritumoral injections (indicated by dashed vertical lines) with the triple combination of therapeutic antibodies (C57BL/6 and RAG1KO). Isotype control antibodies (Iso Ab) were injected into C57BL/6 mice as a growth control. Tumor volumes were measured using calipers and plotted over twenty days. Using one-way ANOVA with Tukey’s multiple comparisons the following was observed: RAG1KO versus Isotype, p=0.002; C57BL/6 vs RAG1KO, p=0.0004. (B) Mice were injected intraperitoneally with a CD8β depleting antibody (CD8 Dep groups) or an isotype control antibody (Iso Dep groups) 3 days prior to tumor inoculation (Day 0). At Day 3, the tumor-bearing mice were peritumorally injected with combination checkpoint antibodies (Combo Ab) or isotype control antibodies (Isotype Ab) and these injections were repeated on day 6 and day 9 (dashed vertical lines). At day 4, a second dose of anti-CD8β antibody or isotype control was injected intraperitoneally to maintain CD8 cell depletion. Tumor growth was monitored via caliper measurements and plotted against time. A two-way ANOVA comparison between treatment groups showed: Iso Dep/Combo versus all other groups, p<0.05. Both (A, B) are single experiments. ANOVA, analysis of variance.

Combination antibody therapy delivered peritumorally is efficacious at lower doses than intravenous delivery and results in an improved toxicity profile

Immunomodulatory checkpoint inhibitor antibodies (eg, anti-PD-1 antibody) are generally delivered intravenously. We, therefore, compared the efficacy of triple antibody (10 µg per antibody) given peritumorally or intravenously. Triple antibody therapy given intravenously to mice with TC-1 ear tumors resulted in tumor regression in 2/10 mice (figure 3A), and a further 2/10 mice showed delay in tumor growth, suggesting that intravenous delivery of antibody was, at the same total dose as peritumoral administration (10 µg/antibody), less effective at promoting TC-1 tumor clearance. To confirm whether this difference was a consequence of triple antibody dose, the triple combination was administered at 10 µg per antibody peritumorally, or 100 µg per antibody intravenously, to mice bearing TC-1 tumors (figure 3B). The higher intravenous dose proved as effective at eliminating tumor as the lower dose delivered peritumorally (figure 3B). However, the larger intravenous dose was associated with significant adverse effects including raised serum ALT, inflammatory liver infiltrates (figure 3C), kidney damage and inflammation of the small intestine (data not shown; as detailed in necropsy report) that was not seen with the lower peritumoral dose. Together, these data suggest that while a higher dose of intravenous antibodies could limit TC-1 tumor growth to a similar extent as a 10-fold lower dose of peritumoral antibodies, this was associated with severe off-target toxicity.

Figure 3

TC-1 tumor regression is achieved at lower doses with peritumoral antibody delivery than intravenous administration and with minimal toxicity. (A) TC-1 tumors were established for 3 days in the ear before intravenous injection of 10 µg of isotype control (left panel) or combination checkpoint antibody therapy (right panel). Tumor growth was monitored with calipers and plotted against time. Data are pooled from two experiments. (B) TC-1 ear tumors were treated with intravenous (100 µg/Ab) (IV) or peritumoral (10 µg/Ab) (PT) delivery of isotype control (Isotype) or combination checkpoint antibody therapy (Combo). Statistical significance was determined using one way ANOVA with Holm-Sidak’s multiple comparison test: IV isotype versus PT isotype and IV combo 100 µg vs PT combo 10 µg, not significant; Combo treatments versus Isotype treatments, p<0.05. Data are from a single experiment. (C) Blood taken from mice prior to the inoculation of tumor (baseline) or from tumor bearing mice treated intravenously with isotype control antibody or combination checkpoint antibody delivered intravenously (100 µg/Ab) (Combo IV) or peritumorally (10 µg/Ab) (Combo PT) were assessed for serum alanine aminotransferase activity using an ALT ELISA (left panel). Statistical significance was assessed using one-way ANOVA with Tukey’s multiple comparisons test: Combo IV versus all other groups, p<0.05, all other comparisons were not statistically significant. Representative H&E staining of liver tissue taken from mice administered with either isotype (Iso) control antibodies (10 µg/Ab PT or 100 µg/AB IV), or triple combination (Combo) antibody (10 µg/Ab PT or 100 µg Ab IV) (right panels). Images are ×40 magnification. The green highlighted box represents an area of immune infiltrate. ANOVA, analysis of variance.

Local combination therapy drives tumor regression at a distant site and augments systemic immune memory

Patients with keratinocyte cancers often present with multiple simultaneous cancers at different skin sites or with recurrence of cancer. To investigate whether a skin cancer at a second site could be impacted by treatment of the cancer at the primary site, and whether immune memory could be developed after treatment of a primary cancer, TC-1 tumors were inoculated into both ears of mice with combination antibody treatment limited to one tumor site (right ear) (figure 4A) This resulted in tumor clearance in both ears in 5/6 mice over two experiments (one further mouse developed a neck tumor). In one mouse that showed initial regression of tumor in both ears, the tumor subsequently returned in both ears, and the recurrent tumors regressed with further combination antibody treatment, suggesting that the tumor had not become treatment resistant (figure 4A, right panel). Eventually, the left ear tumor began to grow again in this mouse despite two rounds of antibody treatment. Using fluorescently labeled antibody injected peritumorally and IVIS imaging, antibody was seen to distribute to the draining lymph nodes and spleen, although at lower levels, after 24 hours suggesting some systemic leakage of the injected antibody from the tumor site (online supplemental figure 2). This distribution is consistent with published literature using IVIS imaging of intratumoral antibody injection in the dorsal flank.22

Supplemental material

Figure 4

Peritumoral antibody treatment of TC-1 ear tumor results in therapy at distant, untreated tumor and long-term, systemic immune memory. (A) TC-1 tumors were intradermally injected into both mouse ears before peritumoral treatment (three injections spaced 3 days apart—see vertical dashed lines) of the tumor in the right ear (RE). The left ear (LE) was not treated. Tumor volumes in both ears were monitored using digital calipers. Each colored symbol in each experiment represents an individual mouse (closed symbol—RE, open symbol—LE on same mouse). Combination antibody treatment of mice (in the RE) is represented by red symbols while isotype treatment of mice (in the RE) are represented by blue symbols. A combination antibody treated mouse whose ear tumors were regressed but later had regrowth of tumors in both ears (experiment 2 right panel) was treated with three more doses of combination therapy (second set of vertical dashed lines). (B) C57BL/6 that had regressed ear TC-1 tumors and remained tumor-free for 50 days were challenged with TC-1 subcutaneously in the flank skin (PT combo treated ear) (left panel). Naive C57BL/6 mice injected with subcutaneous TC-1 without treatment acted as a tumor growth control (Naïve). The tumor volume in the flank was measured and plotted as days post subcutaneous tumor inoculation. Data are pooled from two experiments. Peritumorally, combination treated mice that had regressed ear tumor and been protected against subcutaneous TC-1 tumor rechallenge were held for 150 days (measured from ear tumor inoculation) before the experiment was terminated and splenocytes were stimulated with an MHC-I restricted, HPV16 E7-specific peptide (expressed within TC-1 tumors), for 24 hours and IFN-γ production was detected using the ELISPOT assay (right panel). Splenocytes from a naïve mouse were used as a negative control or individual mice (numbered) treated with the antibody combination with mean and standard deviation shown for assay replicates. Data are from a single experiment and statistical significance was calculated using a Welsh ANOVA test. ANOVA, analysis of variance.

To investigate whether successful triple therapy treatment of tumors generated immune memory, we rechallenged mice that had successfully cleared tumors in a single ear after treatment, and tumor naïve mice, with TC-1 tumor placed subcutaneously in the back skin. Inoculations of TC-1 tumor cells in tumor naïve mice grew progressively as expected (figure 4B, left panel) whereas inoculations in previously treated mice failed to grow, suggesting a circulating, memory immune response (figure 4B, left panel), and this was confirmed by demonstration of tumor antigen-specific, IFN-γ producing CD8 T cells in the spleen (figure 4B, right panel). Peritumoral treatment of ear tumors thus can lead to the development of long-lived anti-tumor CD8 T cells, associated with prevention of tumor growth at a distal site.

The immediate immunotherapeutic tumor response generated by combination treatment is associated with increased granzyme B+ KLRG1+Eomes+ CD8 T cells and reduced CD4 T cells (including CD4+ FoxP3+ cells) at the tumor site

Having demonstrated a pivotal role for the immune system in the success of our treatment, we characterized the immune response within the tumor site by analyzing CD4 and CD8 T cell immune populations within the tumor at day 12 post-tumor inoculation (figure 5A), a time point at which tumor was being eliminated by combination antibody treatment. The fraction of CD8 T cells among the total T cell population within the tumor was increased with combination antibody treatment, relative to isotype control treatment (figure 5B), and the proportion of CD4 T cells including FoxP3+ T cells (figure 5B,C) was decreased. These results are consistent with a role for CD8 T cells in clearing tumor, while also suggesting that induced reduction in CD4+ FoxP3+ T cells can contribute to treatment efficacy. Among the CD8 population, combination treatment was associated with an increase in granzyme B expressing T cells, consistent with increased cytotoxic potential (figure 5D). Granzyme B-expressing CD8 T cells also included more Eomes+KLRG1+ cells in the combination treated tumors relative to isotype treated tumors (figure 5E). Overall, the phenotypic profile of T cell populations within the treated tumors was consistent with increased effector CD8 T cell function.

Figure 5

Peritumoral, combination antibody therapy is associated with an increased percentage of CD8+ granzyme B+ Eomes+ Klrg1+ T cells and decreased CD4+ and Cd4+Foxp3+ T cells. Mice-bearing TC-1 ear tumors were treated peritumorally with a combination of antibodies or isotype matched controls on three separate occasions spaced 3 days apart. On day 12, single cell suspensions of the ear were stained with a panel of fluorescently labeled antibodies before analysis by flow cytometry. (A) Representative plots demonstrating the gating strategy to distinguish T cells. (B–E) T cell subsets defined by antibodies targeting CD4, CD8, FoxP3, granzyme B, Eomes and KLGR1 were analyzed in isotype control or combination antibody treated tumors. Representative plots are shown on the left and summarized data for individual mice in graphs on right. Data are analyzed using t-tests with at least five mice in all groups in 1–3 replicate experiments. ** p<0.01, **** p<0.0001.

Peritumoral delivery of combination antibody treatment is effective in eliminating B16F10 and 8101-PRO tumor growth

Given the success of peritumoral, combination antibody treatment against TC-1 tumors within the ear, we next tested for broader efficacy against two further, unrelated cutaneous tumor models. B16F10 (a melanoma cell line) and 8101-PRO (a UV-induced skin tumor lacking immunodominant antigen expression) were inoculated into the ear of mice to establish a small tumor that was then treated peritumorally with a combination of antibodies against PD-1, 4-1BB and VISTA. In isotype-treated controls, B16F10 tumor grew progressively in all mice (figure 6A; upper left panel). In combination antibody treated mice however, 11/13 mice were tumor-free at day 25 post-tumor inoculation (figure 6A; upper right panel) showing that the therapeutic effects of the antibody combination delivered peritumorally extended to a melanoma cell line. Similarly, combination antibody prevented 8101-PRO tumor growth in 5/8 animals at day 25 (figure 6B), and one of three animals with continued tumor growth showed slowed growth. Consequently, we have demonstrated that local therapy with antibodies targeting PD-1, 4-1BB and VISTA are effective in eliminating a range of cutaneous tumors, suggesting that peritumoral delivery of our checkpoint antibodies may offer therapy against heterogenous skin tumors.

Figure 6

Peritumoral, combination antibody therapy is effective against different ear tumors. (A) B16F10 melanoma tumor cells or (B) 8101-Pro epithelial tumor cells were inoculated into the ear and combination antibody therapy was commenced peritumorally at day 3. Three doses (10 µg/Ab/dose) of either isotype control antibodies (Isotype PT; upper left panels) or antibody combination (Combo PT; upper right panels) was administered as indicated by vertical dashed lines. Tumor volume was measured over time and ear tumor free mice (out of total mice) are indicated on the graphs. This data are further summarized in Kaplan-Meier survival curves (lower left panels) and representative photographs (end of experiment) of mouse ears in combination treated (combination) or isotype treated (isotype control) mice are provided in the lower right panels. Data are pooled from two experiments and statistical significance was determined using a log-rank test (Mantel-Cox).


Our current study addressed two important clinical needs in skin cancer: (1) development of a safe delivery method for combination antibody treatment in primary skin tumors and (2) identification of a novel combination of checkpoint antibodies inducing potent antitumor immunity including systemic memory cells. Given the accessibility of skin cancers, we have demonstrated that local, peritumoral administration of small doses of a triple combination of antibodies targeting PD-1, 4-1BB and VISTA leads to long term, cutaneous tumor regression with systemic immune memory and no observed toxicity in the liver. Antibody combinations involving the targeting of 4-1BB and PD-1 have stalled in clinical trials due to hepatoxicity. This study provides a novel disease indication in which therapeutically relevant doses of previously toxic antibody combinations can treat disease, potentially restarting a process to getting regulatory approval for the safe use of these drugs in skin cancers.

A key feature of our combination immunotherapy was the peritumoral administration of antibodies. Therapies targeting multiple immune checkpoints are often complicated by the induction of autoimmune T cells or other immune-related adverse events particularly for systemically delivered therapy.23 Intravenous administration of immune checkpoint therapies is the gold standard in the clinic for many tumors but our study challenges this paradigm suggesting that local and systemic immune responses can be generated in a safe manner by peritumoral antibody administration for skin cancers. It is not clear in our studies if the antibody combination acts on tumor-infiltrated immune cells, immune cells in the draining lymph node (or beyond) or both although it is clearly possible to separate antitumor immune responses from systemic immune pathology in our system. The importance of checkpoint antibodies, such as those targeting PD-1, acting at tumor draining lymph nodes has been recently described.24 Our data also show that peritumoral injection allows for lower doses of therapy to achieve similar tumor efficacy to higher doses of intravenous antibody. It is likely that intravenous delivery at low antibody dose leads to antibody binding to cells within the bloodstream or other organs such that insufficient therapy reaches the cutaneous tumor site. While there are many literature studies of intratumoral/peritumoral delivery of immunotherapeutic reagents to tumors, the delivery of multiple checkpoint antibodies via this route is an emerging area of interest.25–27 Tumor-directed therapy administered systemically is also an alternative strategy and has been successfully used with a bispecific antibody targeting PD-1 and 4-1BB enabling PD-1+ cells within the tumor and draining lymph nodes to be costimulated via 4-1BB.28 29 Consistent with our work, intratumoral delivery of antibodies against CTLA-4 resulted in delayed tumor growth and protection against distant tumors in an MC-38 subcutaneous tumor model with limited systemic side effects.30 While tumor growth was significantly delayed in this model, it suggested that addition of other immune stimulators might enhance the effect of CTLA-4 monotherapy. In a subcutaneous model using MB49 bladder carcinoma, local administration of anti-PD-1 and CTLA-4 antibodies reduced tumor growth more effectively than either antibody alone.31 Our immunotherapeutic strategy resulted in clearance of B16F10 tumors known to be immunologically “cold” due to reduced MHC-I expression.32 Furthermore, our therapy produced delayed growth of 8101-PRO tumor, a cutaneous tumor cell line derived after chronic UV exposure of C57Bl/6 mice and lacking an immunodominant antigen.17 To our knowledge, this is the first report of an immunotherapeutic immune response to 8101-PRO tumor.

Consistent with previous tumor studies using the TC-1 model, CD8+ cells were necessary for tumor clearance.33 34 Tumor therapy was associated with high Granzyme B levels in CD8 T cells, Eomes+KLRG1+ CD8 T cells and a high CD8/FoxP3 ratio was observed. All of these properties have been associated with either anti-4-1BB antibody therapy in melanoma or tumor survival in the literature.35 36 In addition, tumor antigen-specific CD8 T cells were detectable at day 12 in the TC-1 tumor system consistent with the role for CD8+ cells in tumor therapy. Importantly, local delivery also led to long lasting, systemic antitumor immune responses evidenced in our study by tumor antigen-specific CD8 T cells in the spleen at day 150 after peritumoral treatment and protection against tumor rechallenge at distant sites. In our experiments with concurrent ear tumors, only one of which is treated, it is difficult to determine if therapeutic effects in the untreated tumor was due to antibody leakage or a mobile immune response that had spread from the treated ear. Certainly, we have shown that peritumoral antibody can become systemic (online supplemental figure 2), although at lower concentrations, but whether it accumulates at a sufficient dose to effect therapy in the untreated ear is unknown. One future approach to address the effects of antibody leakage would be to use two different tumors (with no common tumor antigens but susceptible to our combination therapy) in each ear, only one of which is treated with antibody. This is an important issue in cutaneous SCC where susceptible individuals frequently have multiple, simultaneous tumors or have reoccurrence of the tumor over time. With respect to tumor reoccurrence, our studies suggest that local delivery can provide protection against distant, homologous tumors at a later time although again we do not address the significant issue of tumor heterogeneity and whether induced immune responses against one tumor are capable of protecting against a genetically distinct, untreated tumor at a distant point in time. In the future, it will be important to determine if combination antibody treatment elicits CD8 T cell immunity against shared tumor antigens to protect against heterologous tumors.

While antibodies against PD-1 are currently approved for clinical use in metastatic and locally advanced cutaneous SCC,37 combination with antibodies against 4-1BB and VISTA is less well studied. Both PD-1 and VISTA act predominantly as inhibitory molecules for T cell activation in a non-redundant fashion.38 Blocking VISTA interactions alone in tumors using inhibitory antibodies has led to delays in tumor growth, while combination with anti-PD-1/CTLA-4 antibodies improves immunotherapeutic outcomes.39 40 VISTA can contribute to resistance to PD-1 checkpoint immunotherapy and plays a role in skin inflammation.41 42 In our study, both PD-1 and VISTA targeted antibodies as monotherapy cleared tumors in a limited number of individual mice but the underlying cause of this variability in tumor response was not determined and may relate to small differences in tumor size at the time of first therapy. Anti-PD-1 antibody therapy also led to delays in tumor growth consistent with previous studies using subcutaneous TC-1 tumor and intraperitoneal delivery of antibody.43 44 As a monotherapy, the dominant antitumor effect seen in the TC-1 ear tumor model was with agonist antibody targeting 4-1BB. While antibodies against 4-1BB are known to costimulate effector T cells, the efficacy of this monotherapy may also have been due to regulatory T cell depletion within the tumor, as evidenced in our work by decreased FoxP3 cells after treatment, or anti-4-1BB antibody effects on other immune cells such as DC, macrophages or NK cells.45–47 Studies targeting 4-1BB alone in keratinocyte cancers are limited but successful immunotherapy was been achieved in melanoma.48 One study using subcutaneous TC-1 tumors demonstrated a delayed tumor growth in some mice with 4-1BB monotherapy (intraperitoneal delivery) with significant impact on tumor growth when combined with a vaccine.49 A combination of antibodies against 4-1BB and PD-1 delivered intraperitoneally was also shown to be effective in treating subcutaneous B16F10 and MC38 tumors although hepatotoxicity was observed.50 Consequently, the success of our 4-1BB monotherapy may be related to the tumor site (cutaneous in the ear), tumor size at treatment or high local concentrations of antibody after peritumoral injection. Together, the combination of anti-PD-1, anti-4-1BB and VISTA antibodies was superior to monotherapy in the TC-1 tumor model although we acknowledge the dominant role of antibodies targeting 4-1BB. Reducing the concentration of anti-4-1BB antibody within the triple antibody combination or dissecting the effects of double antibody combinations might more clearly demonstrate the benefits of triple therapy in the TC-1 model. However, given that the triple antibody combination does not provide complete therapy in other tumors such as 8101-PRO, our future approach is to supplement the triple therapy rather than dissect the minimal requirements for therapy of TC-1 tumors alone. Using combinations of checkpoint antibodies in tumor immunotherapy may be important in overcoming tumor resistance to monotherapies into the future. In this regard, it would be interesting to combine peritumoral antibodies against 4-1BB and VISTA with the standard clinical treatment of intravenous antibody against PD-1 or PD-L1 for cutaneous cancers.

In conclusion, our study has highlighted the effectiveness and safety of peritumoral delivery of checkpoint antibodies in the therapy of skin tumors, particularly when a unique combination of antibodies targeting PD-1, 4-1BB and VISTA is used. This provides novel preclinical evidence to support a paradigm shift from intravenous to peritumoral combination antibody delivery for safe immunotherapy of primary tumors in human cSCC patients with the potential for protection against cancer recurrence at the same or distant skin sites.

Data availability statement

Data are available on reasonable request. The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.

Ethics statements

Patient consent for publication

Ethics approval

All experiments were conducted under approved protocols by the University of Queensland animal ethics committee (approval no. 2019/AE000404).


We would like to acknowledge the technical assistance of David Sester from the TRI flow core facility, Rona Barugahare, Karen Knox and animal technicians from the TRI Biological Research and Brian Tse, Kamil Sokolowski from the TRI preclinical imaging facility.


Supplementary materials


  • IHF and JLGC contributed equally.

  • Contributors QW and DS performed the experiments. QW, DS, GRL, JLGC, JWW and IHF designed the experiments and helped interpret the data. GRL and QW conceived the project and are guarantors of the content. All authors contributed to the writing and/or editing of the manuscript.

  • Funding This work was supported by a grant from Tour de Cure (RSP-206-2020) and a PhD scholarship from the Australian Government Research Training Program.

  • Competing interests No, there are no competing interests.

  • Provenance and peer review Not commissioned; externally peer reviewed.

  • Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.