Analysis of potential biomarkers of response to IL‐12 therapy

Abstract IL‐12 is a proinflammatory cytokine capable of inducing a wide range of effects on both innate and adaptive immune responses. Its stimulatory effects on T cells and NK cells have led to its classification as a potential inducer of antitumor immunity. Clinical trials have been attempting to harness its immune‐stimulating capacity since the 1990s and have had much success despite notable toxicity issues early on. Several methods of IL‐12 delivery have been employed including i.v., s.c., and local administrations as well as plasmid and gene therapies. However, despite differing methods, dosages, and cancer types utilized in these clinical trials, there are still many patients who do not respond to IL‐12 therapy. This creates an opportunity for further investigation into the immunologic differences between responding and nonresponding patients in order to better understand the variable efficacy of IL‐12 therapy. This review focuses on a limited collection of IL‐12 clinical trials, which further analyzed these individual subsets and detected biologic variables correlating with differential patient responses. A comprehensive review of these potential biomarkers identified 7 analytes that correlated with beneficial patient responses in 3 or more clinical trials. These were increased levels of IFN‐γ, IP‐10, TNF‐α, MIP‐1α, MIG, and CD4+ and CD8+ T cells, with a decrease in VEGF, bFGF, FoxP3+ T regulatory cells, and M2 macrophages. These potential biomarkers highlight the possibility of identifying immunologic determinants of patient response to IL‐12 therapy to conserve valuable resources and benefit patients.


INTRODUCTION
IL-12 is a heterodimeric cytokine that is capable of inducing a wide range of immune effects. It is a member of the larger IL-12 family, which also includes 3 other cytokines; IL-23, IL-27, and IL-35. IL-12 itself is composed of one 35-kDa alpha-chain subunit (p19 or p35) paired with one 40-kDa beta-chain subunit (p40). 1 It is known as a predominantly proinflamamtory cytokine and can be produced by several different cell types including monocytes, macrophages, dendritic cells, neutrophils, and B cells. 2 IL-12 signaling occurs through the IL-12 receptor (IL-12R), which consists of 2 type I transmembrane glycoprotein chains; ∼100 kDa IL-12Rβ1 and ∼130 kDa IL-12Rβ2. IL-12Rβ2 acts as the signal-transduction component by inducing phosphorylation of JAK-STAT proteins downstream. 3 More specifically, activation of IL-12R causes tyrosine phosphorylation of primarily JAK2 and TYK2, which can then phosphorylate STAT1, STAT3, STAT4, and STAT5. 2,4 Although IL-12R is mainly expressed on activated NK and T cells, it has also been found on dendritic cells and B-cell lines. 5,6 The widespread expression of the IL-12R allows for the pleiotropic effects of IL-12 on both innate and adaptive immune cells. For example, IL-12 stimulation of NK and T cells induces production of IFN-γ through STAT4 activation. 7 IFN-γ can then further activate macrophages, NK cells, and B cells and up-regulate T-bet to promote a Th1 T cell response. 8 Furthermore, the ability of IL-12 to stimulate T cells and NK cells has made it an attractive candidate for overcoming the immunosuppressive microenvironments identified in several disparate types of cancer. [9][10][11] The potential for IL-12-induced antitumor immunity led to its identification as a promising therapeutic agent for the treatment of cancer.
After several successful preclinical models demonstrated the benefit of IL-12 for the treatment of cancer, this cytokine therapy progressed into human clinical trials in the 1990s. One of the first clinical trials utilizing recombinant human (rh) IL-12 was reported by Atkins et al. 12  the 500 ng/kg dose level revealed peripheral blood NK cells expressed significantly higher surface densities of CD2, CD11a, and CD56 3 to 4 days after rhIL-12 infusion. Furthermore, NK cell cytotoxicity toward K562 leukemia target cells was significantly increased in PBMCs isolated 3-7 days after dosing compared with cells isolated pretreatment.
A significant increase in PBMC proliferation induced by immobilized anti-CD3 monoclonal antibody was also detected in vitro in PBMCs isolated after a single injection of 500 ng/kg rhIL-12. 13 These encouraging safety and biologic responses led to a followup phase II trial by the same group that administered rhIL-12 at the MTD of 500 ng/kg by bolus i.v. once daily for 5 consecutive days every 3 weeks. 14 However, severe toxicities occurred with 2 deaths and 12 of the 17 treated patients requiring hospitalization. An investigation was subsequently initiated to determine the cause of these unexpected results and a pause was placed on the use of IL-12 in clinical trials. 15 This action created some hestancy toward the safety of IL-12 therapy.
The investigation determined that the test dose had in fact abrogated toxicity and that its omission in the phase II study had contributed to the observed toxicity of the regimen. 14 Follow-up studies were then conducted by Leonard et al. 14 in 1997 to investigate a potential link between IL-12-induced IFN-γ production and IL-12-induced toxicity in the phase II trial. After determining that there were no substantial changes in the drug products used during the 2 trials, focus shifted toward the change in schedule of IL-12 administration. Complementary experiments were performed in C3H/Hej mice. First, mice were treated with 0.5 or 1.0 μg rmIL-12 for 6 consecutive days without pretreatment. This schema resulted in 100% mortality by day 8 and adverse events such as gastrointestinal toxicity and diarrhea, consistent with those events seen in the phase II trial patients. Pretreatment with a single dose of 0.5 μg rmIL-12 a week prior to consecutive dosing however completely abrogated treatment-associated toxicity and prevented death. Consecutive daily rmIL-12 administration to mice without pretreatment resulted in a marked increase in serum IFN-γ levels that peaked 3-4 days later. Conversely, pretreated mice displayed substantial attenuation of serum IFN-γ levels between days 2 and 4. In order to determine if this phenomenon mirrored what had occurred in phase II clinical trial patients, serum samples collected from patients receiving 500 ng/kg i.v. rhIL-12 in both trials were assayed for IFN-γ levels. Samples from phase II clinical trial patients showed substantially higher mean IFN-γ levels with a peak of approximately 27,000 pg/ml on day 4 compared with those from the phase I trial that had a considerably lower peak of approximately 5000 pg/ml on day 3. TNF-α was not detected in the serum of patients from either trial, ruling out the contribution of this additional IL-12-induced cytokine. Thus, in order to confirm the potential link between increased IFN-γ levels and toxicity issues, mice were treated for six consecutive days with 0.5 μg rmIL-12 in the presence of either an IFN-γ neutralizing antibody or a control antibody.
The mice treated with control antibody developed severe toxicity and exhibited 100% mortality after 7 days while IFN-γ neutralization protected IL-12 treated mice from toxicity with 0% mortality. These results were confirmed by corresponding studies in cynomolgus monkeys and allowed for the determination that IL-12-induced toxicity issues may be overcome by controlling the IFN-γ response. Furthermore, it was determined that the single test-dose prior to consecutive IL-12 treatment was capable of controlling the plasma IFN-γ response and abrogating the severe toxicities that previously hindered clinical use of IL-12. Based on this improved understanding of IL-12 biology and the effect of different treatment schedules, there has been a renewed interest in its therapeutic use. 16 Current approaches to IL-12 therapy capitalize on recent protein engineering advancements and utilize IL-12-based molecules with enhanced localization and retention. 17 These molecules include IL-12 fusion proteins or "immunocytokines," which allow for increased targeting of tumor antigens or extracellular matrix proteins, cell surface-tethered IL-12 molecules, IL-12 molecules fused to binding domains and IL-12 molecules with increased circulatory stability. [18][19][20][21] Additionally, IL-12 molecules capable of targeting specific T cell populations are also being developed to improve antitumor responses. For example, Li et al. 22 fabricated a dual-target immune-nanoparticle encasing IL-12 and expressing CD8 and glypican-3 antibodies on the surface, allowing for targeted CD8+ T cell delivery.
This targeted IL-12 delivery method was found to increase the expansion, activation, and cytotoxicity of CD8+ T cells crucial for antitumor responses.
The focus on IL-12 in this review is a reflection of the renewed interest in this cytokine given the advanced technology that is being used to administer it in a more localized and physiologic manner. Moreover, the present review was conducted due to the increased clinical usage of IL-12 in the setting of cancer therapy. 17 This information has allowed for the identification of several potential correlates of response to IL-12 therapy. These correlates may serve as useful biomarkers allowing for more precise monitoring of patient responses during future IL-12 studies. In addition, they could provide a better understanding of IL-12-induced immunity. This review will go on to summarize the correlates of response that have been identified to date.

CORRELATIVE STUDIES CONDUCTED ON PATIENTS RECEIVING i.v. IL-12
Several methods of IL-12 administration have been employed in an effort to determine the optimal delivery system for patient benefit. There were no responses in any of the 19 patients treated below the MTD of 500 ng/kg. Of the remaining 9 patients, one experienced a partial response at 500 ng/kg and 3 experienced disease stabilization; 2 patients at 500 ng/kg and 1 patient at the 700 ng/kg dosage level.
Plasma levels of IL-12 and IL-10 were measured and PBMCs were assayed for in vitro production of IFN-γ and levels of thymidine incorporation in response to stimulation with IL-12 alone or in combination with IL-2 or IL-15. Further analysis was conducted on specimens from eight of the patients treated at 500 ng/kg and 2 patients treated at 700 ng/kg to quantify IFN-γ, IL15, and IL-18 plasma levels postinjection. Three discernable patterns of IFN-γ induction were found among these 10 patients. These included a type-1 pattern, defined as peak IFN-γ levels rising to approximately 450-1600 pg/ml after the first dose, rising an additional 2-3-fold higher after the second dose, and returning to around 450-1600 pg/ml following the seventh dose; a type-II pattern, defined as peak levels after the first dose averaging 2-fold higher than type-I patients, remaining or increasing another 2fold after the second dose, and decreasing following the seventh dose; and a type-III pattern, defined as a modest peak in levels following the first dose with continuingly decreasing levels following the second and seventh doses. An association was also found between the abil- Intravenous IL-12 has also been given in combination with other therapeutic cytokines including IL-2. In a 2003 trial, Gollob et al. 25 investigated the efficacy of i.v. rhIL-12 in combination with lowdose IL-2 for the treatment of patients with metastatic melanoma, transitional-cell cancer of the bladder, or renal cell cancer. rhIL-12 was given twice weekly i.v. at a dose of either 300 ng/kg or 500 ng/kg in 6-week cycles with IL-2 treatment beginning midway through the first cycle. 300 ng/kg rhIL-12 was combined with s.c. administered IL-2 at 0.5 MU/m 2 for the first dose level and in subsequent dosage levels, rhIL-12 was increased to 500 ng/kg and IL-2 was increased successively from 0.5 to 1.0, 3.0, or 6.0 MU/m 2 . Twenty-four patients in total were divided between these 5 dosage levels resulting in 1 patient with a partial response, 8 patients with disease stabilization, and 15 patients with progressive disease. Several analytes were monitored during the course of treatment including plasma levels of IFN-γ, TNFα, IP-10, IL-2, IL-10, macrophage inflammatory protein (MIP-1α), and soluble (s) Fas ligand (L). Lymphocyte subset levels and cytokine receptor expression were also evaluated by flow cytometry. Overall, disease stabilization for more than one cycle occurred primarily in patients receiving doses of IL-2 capable of augmenting IFN-γ levels when combined with rhIL-12. Moreover, increased infiltration of CD4 + and T cell intracellular antigen (TIA)-1 + CD8 + T cells was detected in the tumors of one patient who had significant regression after the addition of IL-2 treatment. This is noteworthy as TIA-1 is a protein associated with granules in cytotoxic T cells and a decrease in the percentage of TIA-1+ tumor infiltrating leukocytes has been shown to correlate with tumor progression in malignant melanoma. 26 Therefore, this study confirms the potential use of increased IFN-γ levels posttreatment as a correlate of response to IL-12 gene therapy and the potential importance of increased CD4 + and TIA-1 + CD8 + T cell tumor infiltration.
A phase I trial conducted in 2004 analyzed the effect of i.v. administered rhIL-12 in combination with the HER2 protein-binding monoclonal antibody trastuzumab in patients with metastatic HER2positive malignancies (breast, n = 12; pancreatic cancer-2; cervical cancer-1). 27 In this study, Parihar et al. 27  In 2009, Bekaii-Sab et al. 28 explored the combination of i.v. IL-12, trastuzumab, and paclitaxel in an attempt to improve treatment efficacy in HER2 oncogene overexpressing cancers. This phase I clinical trial included patients with metastatic HER2-overexpressing cancers (breast, n = 7; colon-6; esophagus-4; stomach-2; pancreas-1; thyroid-1). Trastuzumab was administered i.v. at 4 mg/kg for the first dose and then at 2 mg/kg for subsequent doses on day 1 of each weekly cycle. Beginning in cycle 2, IL-12 was given on days 2 and 5 to dose escalated cohorts at 100 ng/kg i.v., 300 ng/kg i.v. or 200 ng/kg s.c. for cohorts 1-3, respectively. Last, paclitaxel was given i.v. at 175 mg/m2 every 3 weeks. Patients responded well to this combination treatment as there was 1 complete response, 4 partial responses, and 6 patients with stable disease lasting more than 3 months out of a total of 21 patients. In order to identify possible correlates of response, patient plasma was analyzed for cytokine, chemokine and antiangiogenic fac-  29 Intravenous rhIL-12 was given to 11 patients daily at 250 ng/kg for 5 days every 3 weeks and s.c. rhIL-12 was given twice weekly at 500 ng/kg to 31 patients. After completion of the study, 29 NHL patients were evaluable for response with 2 patients having experienced a complete response, 4 patients having partial responses, 10 patients having stable disease, and 13 patients having progressive disease. Additionally, 10 HD patients were evaluable for response with 5 having stable disease and 5 having progressive disease. It is notable that 40% of patients who received rhIL-12 i.v. had a partial or complete response while just 7% of patients who received rhIL-12 s.c. exhibited a clinical response. Both NHL and HD patients were then further evaluated for posttreatment changes in peripheral blood lymphocyte counts by flow cytometry and serum levels of vascular endothelial growth factor (VEGF), bFGF, and IFN-γ by ELISA. Median cell counts of CD4 + T cells were slightly increased from 339 to 342/μl following IL-12 treatment while median cell counts of CD8 + T cells were significantly increased from 423/μl to 576/μl after treatment. The median time to peak CD8 + T cell counts in these patients was found to be 36 days. Moreover, levels of VEGF and bFGF were found to have increased in multiple patients with progressive disease after treatment with rhIL-12. Baseline concentrations of VEGF and bFGF were also higher in those patients with aggressive NHL compared with those with indolent disease. Thus, lower pretreatment levels as well as reduced posttreatment levels of VEGF and bFGF may be potential markers of positive response to IL-12 therapy.

CORRELATIVE STUDIES CONDUCTED ON PATIENTS RECEIVING s.c. IL-12
The s.c. route has also been utilized in several IL-12 studies. Subcutaneous delivery of IL-12 was employed by Alatrash et al. 30   However, only responding patients were found to have a treatmentrelated increase in PD-L1 expression with a concomitant increase in intratumoral mRNA IRF-1 levels following treatment. The ratio of tumor infiltrating CD8 + T cells to Tregs and M2 macrophages was also elevated only in those patients who experienced a complete or partial response. Furthermore, these responding patients had significantly higher intratumoral expression of PD-L1 posttreatment than nonresponders. Conversely, nonresponding patients had higher levels of FoxP3 + Tregs and reduced numbers of tumor infiltrating CD8 + T cells. FoxP3 + Tregs in nonresponding patients were also determined to be in closer proximity to CD8 + T cells than Tregs in responding patients by quantitative spatial analysis of multispectral IHC staining. These findings provide evidence that immune infiltrate composition and configuration can be used as a correlate for the clinical response to IL-12 therapy.

CORRELATES OF RESPONSE FOLLOWING IL-12 PLASMID/GENE THERAPY
In 2020, Bhatia et al. 37 (20), Lenzi (25), Thaker (31) to be highest in patients with progressive disease as posttreatment peritoneal and plasma levels increased by 1.7-fold and 2.6-fold, respec- Serum levels of IL-12 and IFN-γ increased proportionally to the dosage of VDX given and returned to baseline after discontinuation of VDX.
Peak serum levels of IL-12 ranged from 25 to 109 pg/ml and peak IFN-γ levels ranged from 15 to 168 pg/ml across all four cohorts. A positive correlation was found between patient overall survival rate and percentage change in the ratio of CD8 + to FoxP3 + peripheral blood T cells on days 14-28 after treatment. Concurrent corticosteroid use in the overall patient population was also found to negatively affect overall survival. Thus, similar to the finding by the Algazi et al., the peripheral blood CD8 + to FoxP3 + T cell ratio may be considered a useful measure of response to IL-12 therapy.

CONCLUDING REMARKS
The ability of IL-12 to stimulate both innate and adaptive immune responses makes it a powerful cytokine for immunomodulation. It has the ability to promote T cell and NK cell cytotoxic activity and IFNγ production and exerts positive effects on myeloid cells and B cells.
While the longstanding use of IL-12 in the cancer field has generated both concerning and promising results, the improved ability to administer and control IL-12 dosing over time has reconfirmed the potential efficacy of this cytokine therapy. The expanded use of IL-12 therapy has created additional opportunities for correlative analyses that may help to explain the antitumor mechanism of IL-12 and differential patient responses.
In an effort to improve this understanding, several clinical trials have collected various patient samples before and after IL-12 treatment in order to analyze biologic variances between response groups.
In this review, a collection of relevant trials was constructed to document these findings and gauge their potential to function as biomarkers of response to IL-12 therapy. Seven potential biomarkers of response were identified in this review after being detected in at least three separate clinical trials and these are summarized in Table 1. The first biomarker identified was increased levels of IFN-γ in plasma, serum, peritoneal fluid, or RNA transcripts following treatment with IL-12. The fifth and sixth biomarkers identified were similarly increased levels of the chemokines MIP-1α and MIG following IL-12 therapy.
Circulating levels of MIP-1α were increased in clinically benefiting patients in 4 clinical trials, while increased MIG was detected in clinically benefiting patient plasma and serum in 3 clinical trials. MIP-1α is a chemokine that can be secreted by a variety of cell types at sites of inflammation. MIP-1α induces chemotaxis of leukocytes, especially T lymphocytes, which are crucial for the inflammatory response. 44 While there is literature suggesting MIP-1α may also play a role in tumor cell migration, its proimmune effects may be dominant in the context of IL-12 administration. In support of this idea, Jeong et al. 45 found that combining i.v. MIP-1α therapy with preestablished methods of treatment for hepatocellular carcinoma, such as irradiation or sorafenib, significantly enhanced antitumor immunity dependent on increased CD8 + , CD107A + , and CD11C + cells. MIG is a chemokine that like IP-10 is known to be induced by IFN-γ. MIG plays a critical role in chemotaxis, uniquely targeting activated T cells. 46 In previous studies by Park et al. and Tannenbaum et al., 47,48 MIG, along with IP-10, was found to be crucial for IL-12-mediated antitumor immunity. This supports the findings that MIP-1α, MIG and the previously mentioned IP-10 are all chemokines that may function as biomarkers of the multicell response to IL-12 therapy.
The last potential biomarker identified in this review was the decreased RNA transcript and serum levels of the angiogenesis factors VEGF and bFGF. This effect was consistently identified in responding patients in 3 different clinical trials. VEGF is known to be an inducer of tumor angiogenesis as is bFGF. 49 It has also been shown that increasing angiogenesis directly influences tumor volumetric growth rate, which can be largely attributed to increased secretion of VEGF and bFGF. 50 Thus, it is understandable how decreased levels of these factors may lead to reduced angiogenesis and thus better tumor control. 51 Future studies may also benefit from measuring the induction of negative feedback loops resulting from exogenous IL-12 treatment. For example, the p40 subunit of IL-12 can be produced as either a monomer or homodimer and has been shown to antagonize the effects of IL-12 in both mice and humans. 8,52,53 However, p40 has also been shown to promote macrophage chemoattraction, migration of stimulated dendritic cells, and production of IFN-y by CD8 + T cells in mice. 8,54 Thus, p40 induction can exert pleiotropic effects and may be of significant interest to evaluate in response to IL-12 therapy. In this manner, it may be possible to further delineate the antitumor mechanism of IL-12 cytokine therapy and devise ways to make it more effective in a larger proportion of cancer patients.
While this review has highlighted several potential and plausible biomarkers of response to IL-12 therapy, it cannot be said that these findings are conclusive. Additional studies will need to be carried out in order to substantiate these correlates as true biomarkers of response.
Confirmational studies should be conducted in future trials of IL-12 therapy, and exploratory analyses using advanced sequencing and histochemical approaches should be applied to circulating and intratumoral immune cells. The broad reliance on plasma cytokine levels in many of the studies is noted and understandable. It is expected that future studies will include a broader panel of biomarkers such as intratumoral immune infiltrate as measured by multiparameter, immune monitoring tools.