Bringing Clarity to the Murky Problem of Cardiac Allograft Vasculopathy

Development of Mouse Models of Cardiac Allograft Vasculopathy

Although most published studies involving preclinical mouse models of cardiac allograft rejection tend to center on the issues of acute rejection and tolerance induction, arguably the more pressing need in the field of clinical transplantation is better understanding of chronic allograft injury. That is, chronic graft injury is a far greater current clinical problem than is early graft loss due to acute rejection. In heart transplantation, cardiac allograft vasculopathy (CAV) is a form of pronounced coronary artery disease that occurs over time after transplant¹ and remains the major source of transplant loss and recipient mortality.^2,3 Moreover, it has become increasingly apparent over the past several years that the development of donor-specific antibodies (DSAs) is strongly associated with CAV and chronic allograft rejection in general through the process of antibody-mediated rejection (AMR).^4,5 More importantly, once DSA is generated and CAV occurs, limited treatment options are available for preventing eventual allograft failure. As such, in addition to advancing the study of clinical acute and chronic allograft rejection, it is imperative to develop preclinical animal models that permit greater mechanistic insights into the pathogenesis of CAV. Identifying suitable mouse models of CAV, especially involving AMR, has been challenging because useful mouse models need to avoid primary acute rejection and yet be permissive for generating longer-term chronic allograft injury. Early mouse models used to generate CAV employed limited antigen disparity between donor and recipient, such as responses to the minor male H-Y antigen in female recipients⁶ or to donors that differ from the host at a single major histocompatibility complex class II molecule.7, 8, 9 However, these models tended to largely invoke a cellular form of chronic rejection rather than clearly reflect DSA-associated CAV. Another means of developing CAV in mice is through the treatment of recipients with transient or suboptimal immune-modifying agents, which results in chronic rather than acute rejection, although the role of DSA in these models is less defined.9, 10, 11 A more recent approach to assess the role of DSA in triggering CAV is through the direct transfer of antibodies specific for donor major histocompatibility complex class I to immune-deficient recipients bearing a cardiac allograft.^12,13 While these varied approaches have proven useful for studying some aspects of CAV in mice, there remains an onging need to develop improved animal models of DSA-associated CAV.

In the current issue of The American Journal of Pathology, Tsuda et al¹¹ have used a novel and intriguing model of CAV that involves a pronounced DSA response in the allograft recipient. These investigators expand on their prior studies using knockout mice deficient in the chemokine receptor CCR5 (CCR5^−/−) to study AMR in mice.^14,15 Importantly, CCR5^−/− mice develop an exaggerated antibody response after transplant relative to wild-type animals and form a platform for studying the impact of DSA on acute and chronic cardiac allograft rejection. Given that cardiac allograft rejection in this model is CD8 T cell independent,^14,15 this group narrowed the variables of this model by using CCR5^−/−CD8^−/− double-deficient mice as allograft recipients. They then molded this mouse model to reflect chronic AMR by giving transient anti-CD4 therapy that prevents the initial humoral response but permits gradual DSA development and ensuing CAV. Interestingly, this model is strongly biased toward a model of AMR in that there is little T-cell infiltration in the allograft but abundant antibody deposition, characteristics that tend to be uncommon in mouse models of chronic rejection. They further demonstrated that host B-cell depletion with anti-CD20 monoclonal antibody therapy inhibited CAV, strengthening the causal connection between the emergence of DSA and the progression of CAV. Moreover, this model is somewhat variable in that DSA titers differ between individual anti–CD4-treated mice. Rather than this variability being a liability, it is an asset in this case. The authors find that the level of the DSA titer in a given recipient strongly correlates with the degree of CAV, a result also found in pediatric heart allograft recipients.¹⁶ Overall, although this is clearly a somewhat esoteric mouse model, it nevertheless forms an interesting foundation for studying DSA-associated CAV in immune-competent mice.

Using the Molecular Microscope for Understanding the Pathogenesis of CAV

Beyond developing this unique model of CAV, Tsuda et al¹¹ also performed detailed microdissection of cardiac allograft arteries and concomitant transcriptomic analyses of tissue specimens. Within the molecular signature acquired through these studies, the authors find a predominant up-regulation of transcripts encoding delta-like canonical Notch ligand 4 (Dll4), a Notch ligand found in arterial endothelium¹⁷ that plays an important role in angiogenesis.¹⁸ This finding is intriguing given the potential contribution of Dll4 in the vascular remodeling in CAV. Dll4 plays an additional role of inciting macrophage activation by endothelium¹⁹ and therefore could also contribute to local tissue inflammation. Significantly, Dll4 is also up-regulated in clinical biopsies from cardiac allografts undergoing AMR,¹⁹ corroborating the current study by Tsuda et al.¹¹ Furthermore, this study is an excellent illustration of the potential utility of molecular phenotyping in understanding tissue injury. The observed up-regulation of Dll4 in both mouse and human cardiac allografts undergoing AMR makes this molecule attractive for further study as both a diagnostic marker of tissue injury and as a potential therapeutic target for attenuating CAV.

Another result of several basic mouse studies and transcriptomic profiling of grafts undergoing AMR (commonly associated with CAV) is the suggestion that DSA is likely to be insufficient for inducing CAV but rather requires the additional involvement of innate immune cells, especially natural killer (NK) cells and macrophages. In mouse models, results clearly show that CAV triggered by anti-donor major histocompatibility complex class I antibodies requires the active participation and effector function of host NK cells.^13,20 Moreover, Kohei et al²¹ found that NK cells play a key role in AMR of kidney allografts in CCR5^−/−CD8^−/− mice, the same as used in the study by Tsuda et al¹¹ highlighted in this commentary. The potential clinical relevance of NK cells to AMR is supported by findings indicating that NK-associated transcripts are also found in patients with heart or kidney allografts undergoing chronic AMR.22, 23, 24 Thus, a variety of studies continue to directly or indirectly implicate NK cells in the process of CAV and AMR. Although the current study by Tsuda et al¹¹ did not directly test the role of NK cells in their model of CAV, they did find the expression of three NK cell–associated transcripts also found in human allografts undergoing AMR. Taken together, such results obtained both by basic mechanistic studies and through molecular profiling implicate NK cells in contributing to antibody-dependent CAV.

Conclusions and Unanswered Questions: Is There More Than One Pathway Resulting in CAV?

This elegant study by Tsuda et al¹¹ provides an exciting new tool to the repertoire of mouse models for studying CAV and chronic antibody-associated allograft injury in general. However, although the current study by Tsuda et al¹¹ analyzes the features of one model of CAV, it is not clear how well such results can be generalized to encompass the potential range of host responses that result in CAV. This is not so much a criticism of their model, but rather reflects our current ambiguity in defining and understanding chronic allograft rejection.²⁵ It is notable that the clinical histologic profile of CAV can be somewhat variable.³ Thus, it is currently unclear whether common or diverse pathways are ultimately responsible for inflicting CAV. For example, studies described above implicated NK cells as important participants in both mouse and human CAV associated with DSA. However, macrophages are another important innate cell type that have long been associated with chronic allograft rejection.²⁶ More importantly, preclinical studies have made strong causal links between macrophages and the generation of CAV. For example, in some cases, depletion of macrophages can inhibit CAV in mice.²⁷ Additionally, an important study by Zhao et al⁹ found that M2-like macrophages were essential for triggering tissue injury in a different model of CAV in mice. More recently, miRNAs associated with macrophage activity have been targeted to attenuate CAV.²⁸ Although it is unclear whether these differing models of macrophage-dependent CAV also require DSA, such results illustrate the potential for varied routes of host reactivity that may converge to result in CAV.

An important related unanswered question is whether innate cells are always required for antibody-dependent CAV and/or whether differing types of innate cells must interact with each other to mediate CAV. For example, at present, it is difficult to determine whether NK cells or macrophages represent alternative innate pathways capable of triggering CAV or whether they act in concert.²⁵ This is important to resolve because effective therapeutic intervention to attenuate CAV will depend on targeting the appropriate cellular and humoral pathways responsible for mediating tissue injury. Also, the relevance and utility of any given animal model for studying CAV will depend on how well the model reflects a response of unclear diversity in clinical transplant recipients. There is a tendency to equate DSA with chronic allograft rejection, including CAV, subtly implying that there is a defined route of vascular injury. However, although there is a strong correlation between DSA and chronic graft injury, this is not always the case. That is, results from clinical monitoring studies suggest that chronic rejection can sometimes occur without demonstrable detection of DSA and, conversely, the presence of DSA does not always result in chronic rejection.²⁵ Although this could be an issue of detection sensitivity of both DSA and graft histologic assessment, it could also be indicative of alternate mechanisms of CAV. Overall, although the study by Tsuda et al¹¹ is interesting and may be an excellent model for enhancing our understanding of CAV, it cannot address the fundamental issue of how broad the range of host responses might be that can result in CAV.