Universal cancer vaccine

Abstract

Among the potential cancer immunotherapies, vaccination against antigens expressed by endothelial cells lining the tumor vasculature represents one of the most attractive options because this approach may prevent the growth of any solid tumor. Therefore, endothelial cells can be used as a source of antigens for developing a so-called “universal” cancer vaccine. Unfortunately, efficient endothelial cell-based cancer vaccines have not yet been developed because previous approaches utilized direct endothelial cell immunizations which is not effective and can result in the elicitation of autoimmune responses associated with systemic autoimmune vasculitis. Recently, the heterogeneity of the endothelial cell surface was defined using an in vitro system as a means of developing antiangiogenic cancer vaccines. This analysis demonstrated that tumors induced specific changes to the microvascular of human endothelial cell (HMEC) surface thereby providing a basis for the design of endothelial cell-based vaccines that directly target the tumor endothelium.¹ This commentary further describes HMEC heterogeneity from the perspective of designing an endothelial cell-based universal (for the treatment of all solid tumors) cancer vaccine with high immunogenicity that does not pose the risk of eliciting autoimmunity.

Keywords: :

• antiangiogenic cancer vaccine
• universal cancer vaccine
• microvascular endothelial cells
• cell surface profiling
• cell heterogeneity
• vaccine design
• cell proteomic footprinting

Immunotherapies that target the tumor vasculature represent one way of preventing solid tumor growth and metastasis. Included among the various approaches used to elicit immunity against the tumor vasculature is active immunization using endothelial cells. This approach represents the most promising among immunotherapies targeting specific epitopes since cell-based vaccines can target multiple native antigens (most which have not been identified)² and several reports have described the efficacy of this approach.^3-8

However, elicitation of autoimmune responses as a result of anti-cancer vaccination has been demonstrated in both animal models and clinical trials^9-14 and an animal model for autoimmune vasculitis was developed using endothelial cell immunization.^15,16 Autoimmune-mediated damage to microvessels (the primary targets of anti-cancer endothelial cell-based vaccination strategies) leads to systemic autoimmune damage of vessels including the destruction of the vasculature resulting in internal hemorrhage and destruction of internal organs. Therefore, vaccine specificity needs to be considered before an endothelial cell-based vaccine can be developed, that is, vaccines with antigen compositions distinct from antigens expressed by endothelial cells in normal tissues need to be designed as a means of preventing the elicitation of autoimmune responses.

Recently we described the interactions between tumor-induced endothelial cell surface heterogeneity and endothelial cell escape from cell-mediated immune responses.¹ This report presents additional hypotheses that may be applied to the design of a universal cancer vaccine (UCV).

The antigenic composition of a universal endothelial cell-based cancer vaccine

Design and development of cell-based vaccines focuses on the elicitation of immune responses against target cells expressing native antigens.^17-19 Even though whole cells possess a set of cell surface antigens (the set of prioritized antigens for vaccine design^20,21) whole cells also express abundant amounts of intracellular antigens ubiquitous to all mammalian cells that could elicit autoimmunity.²²

Fortunately, immune access to cell surface targets (e.g., by antibodies and cytotoxic cells) suggests that these targets would also similarly be accessible to proteases whose byproducts can be isolated following in vitro proteolytic cleavage.^23-25 Recently, it was shown that the composition of proteolytically-cleaved cell surface targets directly defined target cell killing rates in cytotoxicity assays (CTA) that represent an in vitro anti-cancer vaccination model.^25,26 Therefore, a set of cell surface targets represents the 'antigenic essence' of cells and therefore should be represented by cell-based vaccines. Consequently, targeting immune responses against antigens associated with the tumor vasculature (rather than to antigens associated with the vasculature of normal tissues) should be based on expression differences between endothelial cell surface antigens and the profile present in the tumor vasculature. To identify tumor-associated antigens proteomic footprinting analysis^25,26 was utilized to define the proteolytically-cleaved cell-surface targets. The resulting cell proteomic footprints (CPF) represent a 'snapshot' of HMEC surface targets that are different in tumors and that in CTA directly affects HMEC killing rates following tumor stimulation. This observation provided the basis for developing a UCV comprised of proteolytically-cleaved HMEC surface targets as a means of eliciting immunity against tumor vasculature-associated antigens.

Universality of the endothelial cell-based vaccine

In the context of designing UCV, the most promising finding in vitro was the tumor-type-independent influence of the tumor on HMEC surface antigen heterogeneity (i.e., the tumor influence is not specific to the tumor type, and HMEC heterogeneity is only a result of differences in signal strength). We hypothesized that the in vivo HMEC surface profile was affected by the tumor in the same manner and that the tumor dictated the HMEC antigenic profile as a consequence of the growth stimuli strength. Due to diminishing growth stimuli strength with a distance from the tumor cells (i.e., stimuli of different strength are present simultaneously in vivo) it is expected that HMEC with different surface target profiles simultaneously are present in the tumor-related vasculature. Destruction of any type of HMECs at any location of the tumor vasculature should lead to vessel occlusion, therefore all CTA killing of target HMEC can be directly attributed to the tumor-type-independent HMEC vaccine (i.e., to the universal vaccine).

These observations suggested that effective killing of HMEC targets (stimulated to grow in the presence of human prostate adenocarcinoma cells [LNCap]) was the result of immunity elicited in response to antigens expressed on the surface of HMEC following exposure to human hepatocellular carcinoma (HepG2) cells, supporting the in vitro design of the universal vaccine with efficacy of targeting equal to 2.45. Efficacy was defined as the fold difference between the number of killed target HMEC stimulated to grow in the presence of tumor cells vs. the number of killed HMEC targets stimulated to grow in the presence of normal tissue. This efficacy provides a therapeutic window where tumor HMEC cells could be killed before normal tissue HMECs are adversely affected.

Efficacy of UCV in vitro

While the diversity of HMEC targets is present in vivo (due to tumor stimuli of different strengths) and the diversity of antigen compositions can be prepared from HMEC cultures, CTA can describe UCV efficacy in the context of the limited number of experiments with specified cell targets and antigens. Therefore, CTA data were further examined by an approximation that defined the dependence of target killing rates based on the similarity of their surface profiles to the surface profile of cells used to generate antigens used to elicit the immune response measured in the CTA (Supplementary Material and Table 1). These data suggested that an efficient UCV could be generated utilizing the surface antigens of HMEC cultures and that HMEC cultures stimulated to grow in the presence of the tumor should develop a CPF with a correlation equal to 0.82. In this scenario, the efficacy of the UCV will exceed 18 (i.e., 18 tumor endothelial cells will be destroyed before 1 normal endothelial cell is destroyed).

Table 1. Efficacy of target HMEC killing in CTA calculated by approximation of the experimental data

	Δr
		0.01	0.03	0.05	0.1	0.2
r	0.55	3.68^a	3.67	3.65	3.61	3.52
	0.6	1.34	1.3	1.27	1.18	1.01
	0.7	1.22	1.15	1.09	0.93	0.61
	0.75	1.72	1.61	1.5	1.23	0.69
	0.77	2.21	2.07	1.92	1.55	0.82
	0.78	2.63	2.45	2.27	1.83	0.93
	0.79	3.28	3.06	2.83	2.26	1.11
	0.8	4.44	4.13	3.81	3.02	1.44
	0.81	7.04	6.53	6.02	4.74	2.19
	0.82^b	18.03	16.7	15.36	12.03	5.36

^a Efficacy of HMEC targeting (efficacy of UCV in vitro) in CTA expressed as the ratio of the number specifically lysed target HMEC stimulated to growth in the presence of tumor cells over the number lysed following culture in the presence of normal tissue. ^bBased on the HMEC specific lysis estimate, lysis should not be observed following growth in the presence of normal tissue at r equal to 0.83. Therefore, the efficacy of HMEC targeting in a specific CTA can not be calculated (tending toward +∞) at r ≥ 0.83. r – correlation between CPF of two HMEC cultures, one of which is used as a source of autologous antigens in the CTA for which the UCV efficacy is approximated. ∆r – absolute differences between 2 correlation values. The first correlation is calculated between the CPF of HMEC cultures (one which is used as the source of autologous antigens) and the second correlation is calculated between the CPF of HMEC cultures (one of which is used as autologous target in the CTA) for which the UCV efficacy is approximated.

It should be noted that these UCV efficacy reference values are directly affected by conditions associated with the CTA and the proteomic footprinting protocols used. Moreover, due to the autologous nature of the vaccine, the cellular proteomic footprint should be established for each individual patient’s HMEC that is to be vaccinated.

Transitioning these observations into clinical practice will require that the premise of the UVC vaccination strategy be understood. Specifically, that the HMEC culture used for antigen preparation be sufficiently affected by culture in the presence of the tumor to yield a formulation that is both safe and effective. The following steps can be employed in the UCV design:

• Establish a correlation between 2 HMEC primary cultures (from 2 patients) by defining their respective CPF. Secondary passages should be avoided.

• Both HMEC cultures should be affected in similar manner by tumor-conditioned medium, that is, their surface profiles should be similarly changed. In addition, their CPF should be measured again since CPF similarity is a direct reflection of the degree of change to the cell surface profile following tumor stimulation.

• Establish a correlation between the efficacy of HMEC targeting (in the CTA) and CPF correlation values. The correlation value for the most efficient HMEC targeting should be calculated and autologous target cells from the person receiving the vaccine should be used in the CTA.

• Tumor-conditioned medium containing different concentrations and/or different growth factor compositions should be used to provide the required growth stimuli for 2 HMEC cultures to reach the required CPF correlation value. Even if an ideal correlation value is not reached, the actual CPF correlation will show the efficacy of UCV prepared using the obtained antigen composition. If the efficacy is acceptable, autologous HMEC surface targets should by used for vaccination (with adjuvant).

It is important to consider that HMEC surface targets intended for vaccine development (as well as for cell proteomic footprinting analyses) be obtained from HMECs treated with trypsin. HMEC should be live during treatment to avoid possible contamination of collected proteolytically-cleaved HMEC surface targets with undesired intracellular contents released from cells destroyed during treatment. Trypsin with high activity and purity should be used to exclude undesired contamination of cell surface targets intended for analysis (to establish the CPF) and for vaccine preparation.²²

Conclusions and Perspectives

From the in vitro experiments described it can be concluded that a safe and efficacious personal universal vaccine can be developed. The universal vaccine described is comprised of cell surface targets collected from autologous HMEC stimulated in a definite way with tumor cells. The nature of the cell surface targets can be confirmed using cell proteomic footprinting.

Future studies (animal models, adjuvant selection, and vaccination schedule) for vaccine development are required. However, previous animal and human studies corroborated the capability of in vitro induced specific cytotoxic cells to mediate in vivo protection against tumor challenge.^27,28 These data suggest that a highly effective UCV can be developed.

Sources of support

This work was funded by ZAO BioBohemia (Moscow, Russia).

Disclosure of Potential Conflicts of Interest

Authors declare that they have not potential conflicts of interest.