Assays on the Influence of Biomaterials on Allogeneic Rejection in Tissue Engineering

Setting up meaningful in vitro assays

To simulate in vivo implantation of a tissue-engineered construct, it is critical that proper attention be paid to the delivery of the biomaterial in vitro and to the choice of relevant responder cells. Most tissue-engineered constructs include differentiated cells seeded on a large, three-dimensional biomaterial scaffold to which many different cells involved in wound healing and the immune system respond. Determining the critical components of an in vitro experiment requires careful consideration. Methods to analyze the effect of in vitro contact between materials and cells are summarized in Table 1.

Measuring the direct effect of the biomaterial on innate immune cells

The primary effect of the biomaterial component is likely to be on cells of the innate immune system, such as macrophages and dendritic cells. There are a number of ways of measuring the level of activation for biomaterial screening purposes. Some assays to elucidate the mechanism by which biomaterials are recognized will also be described.

Viability and adhesion of macrophages seeded on biomaterial surfaces is sometimes used as a measure of “biocompatibility.” As noted above, the significance of in vitro adhesion on in vivo responses is not altogether clear. Ainslie ³³ used cell survival and adhesion of macrophages on polytetrafluoroethylene nanofibrous surfaces as such an indicator. Adhesion to surfaces is linked to macrophage fusion and subsequent foreign body giant cell formation, ³⁴ which is a concern for regeneration but not for subsequent influences on adaptive immunity. Biomaterials have been shown to induce other activation markers that are important in device design but less likely to influence immune responses. For example, activated U937 cells release degradative enzymes including esterase and acid phosphatase. ³⁵

One simple and commonly used assay measures the release of cytokines by cells incubated with biomaterials. The supernatant is removed at times varying between 12 and 96 h, and the cytokine levels are quantified using enzyme-linked immunosorbent assays (ELISAs). In some cases, intracellular cytokine levels are measured using flow cytometry, or messenger RNA (mRNA) levels are measured using polymerase chain reaction. Cytokines will affect inflammation and will also influence activation levels of lymphocytes. This will potentially compromise the function of the tissue-engineered construct and induce further “danger signals” ³⁶ that will stimulate innate immune responses.

Production of tumor necrosis factor alpha (TNF-α), interleukin-6 (IL-6) and IL-1 beta (IL-1β) is most often monitored. For example, rat alveolar macrophages upregulate TNF-α and IL-6 mRNA in response to a range of biomaterials. ³⁷ Similarly, silicon dioxide nanoparticles induced TNF-α and IL-1β. ³⁸ These three cytokines are a small subset of a much wider range of cytokines that have been analyzed. For example, RAW 264.7 cells were shown to release high levels of TNF-α, granulocyte colony-stimulating factor (CSF), monocyte chemoattractant protein-1, regulated upon activation normal T cell expressed and secreted, granulocyte macrophage CSF, and eotaxin in response to alginate oligosaccharides. ¹² Cytokines promoting Th2 bias, IL-13 and IL-25 were shown to be upregulated in alveolar macrophages in response titanium dioxide particles. ³⁹ Hydroxyapatite particles upregulated IL-18 in human monocytes. ⁴⁰ Other cytokines including IL-1α and IL-9, ¹² interferon-gamma, IL-4, and IL-5 ³³ have been studied but have not been shown to be strongly upregulated by biomaterials.

Flow cytometry of cells exposed to biomaterials displays changes in surface markers. Babensee et al. have shown alterations particularly in the costimulatory molecules CD80 and CD86 and maturation marker CD83 in human and mouse dendritic cells due to biomaterial contact.^41–43 They also studied CD40 and MHC expression, but these were not significantly altered in their experimental system. Such changes would have important consequences for efficiency of antigen presentation to T lymphocytes. Differentiation of dendritic cells has the potential to affect the immunological outcome profoundly; antigen presentation through immature dendritic cells has been associated with production of regulatory T cells.^44,45 Similarly, plasmacytoid dendritic cells have been associated with allotolerance, ⁴⁶ whereas myeloid dendritic cells have been used in immunotherapy against cancer. ⁴⁷ Thus, the dendritic cell phenotype induced by biomaterial contact has the potential to cause accelerated rejection or immune tolerance.

Macrophages and neutrophils will often release reactive oxygen intermediates (ROIs) as a result of inflammatory stimulation. Increased levels of ROI release have been measured in response to biomaterial stimulation. Yamamoto employed a superoxide-reactive chemiluminescence producer and an optical fiber probe to demonstrate ROI production in response to blood dialysis membranes. ⁴⁸ Similarly, macrophages stimulated in vitro with PLGA microspheres showed increased ROI production. ⁴⁹ Among other possible reactions, increased ROI levels will induce local tissue damage that can compromise the function of the cells in the tissue-engineered construct, which would cause subsequent innate-activating danger signals — an inflammatory feedback loop.

The effect of biomaterials on gene expression in innate immune cells has also been studied. The advent of microarray technology means that it is now possible to study a large number of responses simultaneously. There are “inflammatory” microarrays available that narrow the spectrum of targets. Such systems are sensitive to variability in experimental conditions, but with careful selection of controls, in vitro biomaterial testing with microarrays has great promise. Some preliminary work has shown the efficacy of microarrays in highlighting responses to particles. ⁵⁰

In addition to screening the effects of biomaterials on innate immune cells, determining the mechanism of activation is important to design improved biomaterials. It has been hypothesized that some polymeric biomaterials signal through Toll-like receptors (TLRs) on innate immune cells. ¹⁴ Metallic and ceramic nanoparticles do cause upregulation of TLR3, TLR7, and TLR9. ³⁸ One key downstream effect of signalling though many of the TLRs is translocation of the transcription factor nuclear factor-kappa B (NF-κB) from the cytoplasm to the nucleus. ⁵¹ This can be monitored using immunofluorescence or by separating nuclear and cytoplasmic fractions and running Western blots or ELISAs. ⁴¹ Electrophoretic mobility shift assays also show if NF-κB has bound to DNA, ⁵² and there are antibodies that distinguish the active and inactive forms of NF-κB. ⁵³ On its own, NF-κB analysis can be used as a screening tool, but it can also be powerfully combined with the use of cells from receptor knockout sources. It seems that receptors are involved in biomaterial recognition. ⁵⁴ Recently, there has been interesting work showing that antibodies to TLR2 and TLR4 block RAW 264.7 responses to alginate oligosaccharides ¹¹ and that antibodies to TLR4, CD14, and CD3 block the effect of chitooligosaccharides. ¹⁴ Similarly, the inflammatory effect of hydroxyapatite disappears in TLR4 −/− macrophages. ⁵³ Scavenger receptors have also been implicated; the lung inflammatory defense against particles was lower in mice that were −/− for the macrophage receptor with collagenous structure (MARCO) ⁵⁵ and for SR-AI/II −/− mice. ⁵⁶

Measuring the direct effect of the biomaterial on adaptive immune cells

It is generally assumed that biomaterials do not interact directly with cells of the adaptive immune system (e.g., T and B lymphocytes), despite the observation that lymphocytes are attracted to biomaterial implant sites. ⁵⁷ Most biomaterials are unlikely to elicit an antigen-specific response directly in T cells; they do not have MHC molecules, and most biomaterials are not protein based. Notable exceptions are emerging as protein–peptide engineering becomes more developed. Biomaterials based on collagen, elastin, and spider silk (to name just a few) have been proposed.^58–62 Even protein-like biomaterials based on engineered amino acids have been designed. ⁶³ These materials have the potential to activate adaptive immune responses. Additionally, some biomaterials extracted from xenogeneic sources have been used. A biomaterial based on porcine small intestinal submucosa has been shown to cause antibody production, although with no clinically adverse consequences. ⁶⁴ Nonetheless, most biomaterials proposed for tissue engineering are polymeric and not polypeptides.

B cells can recognize a wider range of antigens (including polysaccharides) than T cells, leaving open the possibility that B cells will produce antibodies specific to biomaterials. It is also well known that proteins adhere strongly and rapidly to most polymeric biomaterials. ⁶⁵ Particularly during the breast implant controversy, many researchers proposed that native proteins adhered to biomaterials became rearranged, exposing new epitopes ⁶⁶ that were recognized as antigens, leading to autoimmunity. The lymphocyte activation by such cryptic epitopes is possible but has not been conclusively proven.

Stimulation of lymphocytes could occur through a number of mechanisms. The material itself could stimulate the cells by unknown means. Alternately, proteins bound to the biomaterial surface (e.g., opsonins such as non-specific antibodies that activate innate cells) might cause activation. Some proteins change conformation and become activated as a result of biomaterial contact. For example, biomaterials have been shown to cause formation of complement fragments;^67,68 in turn, such complement fragments have been shown to activate leukocytes ⁶⁹ and even induce regulatory T cells. ⁷⁰ If biomaterials directly activate lymphocytes, this will have important consequences for the reaction to transplanted allogeneic cells. Recent evidence has emerged that T cells also express TLRs ⁷¹ and that ligation of the TLRs on CD4+ cells is involved in regulating regulatory T cells. ⁷²

The most straightforward way to monitor the effect of biomaterials on lymphocytes is simply to co-incubate the biomaterials in vitro with purified lymphocytes. Measuring proliferation, usually through the incorporation of tritiated thymidine or bromodeoxyuridine into the DNA of dividing cells monitors lymphocyte response. Otsuka recently showed that phenylboronic acid residues and surface wettability of polymers affected lymphocyte proliferation in vitro. ⁷³

For T cells in particular, division does not necessarily indicate function. The cytokines they produce influence the outcome. Th1 cells (with signature interferon-gamma (IFN-γ) secretion) are typically associated with rejection, whereas Th2 cells (with archetypal IL-4 production) are controversially linked to a less robust rejection response.^74,75 Newly identified Th17 cells (with IL-17 secretion) are associated with autoimmunity and might be involved in transplant rejection. ⁷⁶ Finally, a range of regulatory T cells (e.g., Foxp+CD4+CD25+ cells, IL-10–secreting Tr1 cells, and transforming growth factor beta (TGF-β) secreting Th3 cells) have been associated with transplant tolerance.^77,78 Tolerance induction is clearly desirable in tissue engineering. Thus it would be informative to measure the cytokines secreted into the supernatant of these cultures. Individual ELISA kits could be used; there are also multiplex beads that measure a panel of T helper cytokines using flow cytometry.

With a slightly different perspective, one interesting study was done that examined the effects of biomaterials on cell differentiation during primary bone marrow cell culture. When a tissue-engineered construct is implanted, it causes wounding, which has in turn been shown to cause recruitment of bone marrow cells.^79,80 When bone marrow cells were cultured with polyethylene or cobalt chromium molybdenum particles, they induced proliferation of CD14+ (monocyte/macrophage) and CD66b (granulocyte) cells, respectively. ⁸¹ Markers for T cells (CD4+), B cells (CD19+), and hematopoietic stem cells (CD34+) were also analyzed.

Measuring the overall effect of the biomaterial on adaptive immune cells

Because biomaterials cause inflammation in vivo, thus activating macrophages (APCs), most researchers in the field believe that influences of biomaterials on adaptive responses will occur indirectly because of activation of APCs. However, simply knowing the level to which a given biomaterial activates an APC in vitro will not predict the effect on the adaptive immune response. Thus it is necessary to measure the influence of biomaterials in an assay more representative of the rejection response.

Biomaterials incubated with a mixture of cells will show the overall response of innate and adaptive immunity, although it will not give information directly about the effect of biomaterials on rejection. Borges et al. incubated unfractionated splenocytes with soluble chitosan and alginate. ⁸² They found that an early cell activation antigen, CD69, was upregulated on B cells and T cells, although proliferation was unaffected. It is not clear whether interactions between innate and adaptive cells influenced the result.

Anderson et al. have shown that interactions between ymphocytes and monocytes influence the response to biomaterials, even in the absence of antigen. Lymphocytes incubated with tissue culture polystyrene and poly(ethylene terephthalate)–based biomaterials increased monocyte attachment, and monocytes incubated with biomaterials increased lymphocyte proliferation in a material-chemistry-dependent manner.^83,84 Direct contact was not required. Such studies illustrate the complex interactions that occur in the foreign body response.

The MLR is the standard in vitro method in transplantation to test for allogeneic rejection. Typically, the stimulator cells are treated using gamma irradiation or mitomycin C to prevent proliferation. ²⁸ For testing tissue-engineered constructs, the stimulator cells will be the cells that would be included in the construct. Responder cells (splenocytes in murine models; PBMCs in human models) are incubated with the stimulator cells, usually for 4 to 5 days, and the resulting proliferation is measured. MLRs are effective tools for measuring allogeneic responses, because up to 10% of T cells will respond directly to MHC mismatches. This means that proliferation in response to primary stimulation is measurable. This technique is not, however, sensitive to primary stimulation with processed antigens presented by recipient APCs (indirect presentation), although it can measure memory responses to indirect stimulation. Indeed, the MLR has been shown not to reflect clinical tolerance.^85,86 This restricts the ways in which this assay can be used in vitro without in vivo priming in measuring indirect effects of biomaterial stimulation.

One can simply add the biomaterials to an MLR and determine the effect on proliferation. Here, it is important to include appropriate controls to determine the effect of the material in the absence of allogeneic stimulation. Babensee et al. has shown that dendritic cells pre-incubated with PLGA and, to a lesser extent, agarose stimulate greater allogeneic T cell proliferation.^41,43 Although an excellent indication of dendritic cell activation, this test reflects direct rather than indirect rejection responses. Depletion of cells from these assays (e.g., by depleting antibodies) will determine the affected cells, and assaying the cytokines in the supernatant will give more information about the type of response (e.g., Th1, Th2, or Tr1).

In vaccine adjuvant research, tools have been developed to further parse responses, even without primary stimulation in vivo. For a number of model protein antigens, researchers have determined the dominant peptides that are present in the MHCI and MHCII binding clefts. Antibodies have been developed that recognize the peptides in the binding cleft. Similarly, many tetramers are commercially available that bind to specific T cell receptors, usually on CD8+ cells. Moreover, T cell receptor transgenic mice have been developed in which all the T cells respond to a specific peptide. For example, OT-I mice are transgenic for the CD8+ T cell receptor, recognizing peptides 257-264 of hen egg ovalbumin (OVA) in the context of H-2 Kb. Also, OT-II transgenic mice express the CD4+ T cell receptor specific for OVA323–339. Because every T cell is responsive, primary MLRs become possible. Using such systems, Hamdy et al. have demonstrated that recipient dendritic cells efficiently process OVA delivered with PLGA nanoparticles and present them to recipient T cells. ⁸⁷ In allogeneic transplantation models, similar experimental tools are not yet available. The vaccine adjuvant work suggests that biomaterials are likely to influence the indirect component of the allogeneic rejection response.

Setting up in vivo assays

In striving to understand innate and adaptive responses to tissue-engineered constructs, in vivo models provide information about complex interactions that are impossible to capture using in vitro assays. The most-accurate outcomes would be measured by testing a final “perfected” version of a tissue-engineered construct in the target location, but to incorporate feedback from immune responses, it is beneficial to create simpler models that provide information about appropriate biomaterial selection before full construct development. There are several key variables to consider in designing in vivo models. First, the model construct geometry should be reflective of the final design. As discussed earlier, the size of the biomaterial features relative to that of the cells is likely to influence the response. Second, the implant location will probably affect the immune respons. ⁸⁸ Immune-privileged sites such as the cornea or testis will give dramatically different responses,^89,90 but more-subtle location choices such as intraperitoneal versus subcutaneous implantation might also alter responses. Third, as with in vitro assays, the selection of cells should reflect the “worst case” immunologically. For example, if a range of cells will be included in a proposed construct, the ones that have the highest MHC expression should be used. In addition, ensuring MHC mismatch between donor cells and recipient is important. Finally, surgical technique is critical; the greater the level of wounding, the higher the “danger signals” to the innate immune system. ³⁶ Careful attention to controls should include naïve animals, sham surgeries, cells alone, and biomaterials alone. Methods to analyze the effect of in vivo implantation of constructs with biomaterials and cells are summarized in Table 2.

Measuring local outcomes

Straight monitoring of the local environment around an implant is informative, although the mechanism is not elucidated, and survival of transplanted cells is an essential indicator. These methods are, however, often too crude to distinguish subtle (but ultimately important) responses to different biomaterials. This review is primarily intended to cover in vitro analysis of responses (with or without implantation) and so will examine local responses and transplant survival in only a cursory manner.

Perpetual survival of the cells in a tissue-engineered construct is the desired outcome. It is a reasonable assumption that, without using anti-rejection drugs, allogeneic cells will not survive in vivo, but real-time monitoring of in vivo cell survival is not easy. There are some models in which the cells perform a function that can easily be monitored, as with islet transplants altering blood glucose levels (e.g., ^91,92). For many models, though, the only method of monitoring function is to sacrifice the animal and look for markers that distinguish donor cells from recipient cells within the construct. Ideally, cell markers (metabolic or visual) will be introduced that allow real-time monitoring of the viability of implanted cells, without sacrificing the animal. Noninvasive monitoring has been accomplished using proton NMR imaged choline production from betaTC3 cells. ⁹³ Magnetic resonance imaging has also been used to followed cells in vivo, where the cells have been labeled with superparamagnetic iron oxide. ⁹⁴ One intriguing study used cells that were transfected such that NF-κB activation caused EGFP production and fluorescence, thus allowing non-invasive monitoring of stress to cells within an implanted construct. ⁹⁵

Histology of explanted tissue-engineered constructs will provide valuable information about cell survival and about recipient immune cells active at the implantation site. If the construct were implanted in the peritoneal cavity, local cells could be harvested using peritoneal lavage. The cells in the lavage can then be analyzed after a cytospin and H&E stain or analyzed using a flow cytometer after the cells are labeled with antibodies. Recipient cells attracted to the implant site can also be analyzed for gene expression profile, including using microarrays.

Ho et al. developed one elegant technique for visualizing responses to implants. ⁹⁶ Briefly, they created a transgenic mouse that had a luciferase reporter gene linked to a NF-κB-responsive promoter. Thus, when NF-κB was stimulated to translocate to the nucleus, the cells emitted light that was monitored using in vivo bioluminescent imaging, allowing ongoing, non-invasive observation of the inflammatory response.

Measuring the effect on adaptive immunity

Evidence is accumulating that biomaterials affect innate immune responses. In turn, it has been hypothesized that adaptive responses will be altered. A challenge in developing assays to test this in an allogeneic setting is that biomaterials will thus affect the indirect component of the rejection response. Although indirect responses (recipient APCs activating recipient T cells) are sufficient to cause rejection,^97,98 the robust primary direct response (recipient CD8+ T cells interacting directly with donor cells) that dominates allogeneic transplantation masks them in many assays. Tolerance responses are primarily induced through indirect activation. One tantalizing possibility in tissue engineering is that, through appropriate biomaterial and cell selection, a tolerant state could be induced, negating the need for a lifetime of anti-rejection drugs. To measure the influence of biomaterials on the indirect response, the construct is implanted (priming the immune system), and subsequent assays are performed. These include assays that examine the effect of T cells, B cells, and dendritic cells.

As described above, the MLR is the classic assay in allogeneic transplantation. When determining the level of primary activation in vivo, recipient splenocytes are re-challenged with donor cells or with allogeneic third-party cells. The secondary response to the donor cells in rejecting recipients is greater than the primary response to the third-party cells.^99,100 Hyporesponsiveness in this assay is correlated with less-robust rejection (e.g., effective immunosuppression) but has not been linked to specific tolerance. ⁸⁵ The biomaterial component of the construct will potentially alter the magnitude of this response. Indeed, a similar assay has been performed in a xenogeneic model, in which indirect rejection dominates the T cell component. There, a biomaterial combination of agarose and a copolymer of hydroxymethylmethacrylate and methylmethacrylate implanted coincident with a xenogeneic skin graft was shown to transiently suppress the second-set MLR proliferative response. ¹⁰¹ This demonstrated that biomaterials are not necessarily activating and deleterious in tissue engineering; it is likely that they will alter immune responses but not necessarily for the worse. In an allogeneic system, the MLR does not distinguish between direct and indirect responses. Direct and primary responses will mask the influence that biomaterials have on the indirect and memory response. ¹⁰²

The enzyme-linked immunosorbent spot (ELISPOT) assay overcomes some of the limitations of the MLR in these circumstances. Responder cells from in vivo primed animals are challenged with stimulator cells at a lower concentration than in the MLR. The reaction occurs over a much shorter time than the MLR (24 h) on a cytokine-coated plate. After incubation, the plate is developed such that each cell that produced the target cytokine causes a spot. Because of the short incubation time, memory responses are more easily distinguished from primary responses. In addition, the T cell subsets elicited using transplantation can be distinguished by using ELISPOTS for cytokines produced by Th1 cells (e.g., IFN-γ, IL-2), Th2 cells (e.g., IL-4, IL-5), or even Tr1 cells (IL-10), although this assay does not work to detect TGF-β-producing Th3 cells. ¹⁰²

The trans-vivo delayed-type hypersensitivity assay highlights the indirect and regulatory components of transplant responses. Developed by VanBuskirk et al. to detect tolerance in human transplant recipients, it takes advantage of memory CD4+-dependent inflammation.^103–107 Recipient PBMCs are mixed with freeze-thawed donor cell lysate and injected into the footpad of a severe combined immunodeficiency mouse. Samples from donor-sensitized patients led to footpad swelling within 24 h. Non-swelling regulatory responses could be distinguished from non-regulatory responses by including a recall antigen; regulatory cells suppressed the swelling response to the recall antigen. Furthermore, by using antibodies to IL-10 and TGF-β, the type of regulatory cell could be determined.

As discussed above, plasmacytoid dendritic cells have been implicated in transplant tolerance. Mazariegos et al. have proposed using the plasmacytoid dendritic cell:myeloid dendritic cell ratio in recipient blood as a measure of tolerance, although this is controversial.^102,108,109 It would be an interesting correlate to direct innate in vitro assays to determine whether tissue-engineered constructs induce this tolerogenic dendritic cell subset in vivo.

Finally, one common method of evaluating rejection responses is to measure MHC alloantibodies in recipient blood. Clinically, this assay is performed routinely in transplant recipients. Anti-HLA antibodies have been linked particularly with chronic rejection. ¹¹⁰ Methods using ELISA and flow cytometry have recently been developed that facilitate such analyses. ¹¹¹ There is some evidence that tolerant patients do not have circulating anti-HLA antibodies. ¹⁰⁸