Hydrodynamic gene delivery to the pig liver via an isolated segment of the inferior vena cava

Hydrodynamic gene delivery is an attractive option for non-viral liver gene therapy, but requires evaluation of efficacy, safety and clinically applicable techniques in large animal models. We have evaluated retrograde delivery of DNA to the whole liver via the isolated segment of inferior vena cava (IVC) draining the hepatic veins. Pigs (18-20 kg weight) were given the pGL3 plasmid via two programmable syringe pumps in parallel. Volumes corresponding to 2% of body weight (360-400 ml) were delivered at 100 ml s(-1) via a Y connector. The IVC segment pressure, portal venous pressure, arterial pressure, electrocardiogram (ECG) and pulse were monitored. Concurrent studies were performed in rats for interspecies comparisons. The hydrodynamic procedure generated intrahepatic vascular pressures of 101-126 mm Hg, which is approximately 4 times higher than in rodents, but levels of gene delivery were approximately 200-fold lower. Suprahepatic IVC clamping caused a fall in arterial pressure, with the development of ECG signs of myocardial ischaemia, but these abnormalities resolved rapidly. The IVC segment approach is a clinically acceptable approach to liver gene therapy. However, it is less effective in pigs than in rodents, possibly because of larger liver size or a less compliant connective tissue framework.

Cardiovascular function following acute volume overload for hydrodynamic gene delivery to the liver

Hydrodynamic gene delivery to the liver is a valuable experimental tool and an attractive option for nonviral gene therapy of liver disease. However, little attention has been paid to the major obstacle to clinical application: acute volume overload of the cardiovascular system. We delivered volumes of DNA solution (pGL3 plasmid) corresponding to 1, 2, 4, 6 and 8% of the body weight at 100 ml/min to the inferior vena cava (IVC) of DA strain rats. Central venous pressure (CVP), arterial pressure, pulse and electrocardiogram (ECG) were continuously recorded for subsequent analysis. Each volume produced a characteristic response, but all (including the 1% volume) caused severe falls in blood pressure and pulse within 1-2 s of the infusion, with ectopic beats and widening of the QRS complex in the ECG. The response to volumes of 4% and higher suggested that the liver acted as a volume sink, mitigating the immediate effects of volume overload. The 6 and 8% volumes caused profound and protracted falls in blood pressure and pulse, with a multitude of severe electrical abnormalities in the heart, including electromechanical dissociation. Vagal blockade with atropine, and the use of Ringer's solution to prevent electrolyte disturbances, did not ameliorate this picture.

A remarkable permeability of canalicular tight junctions might facilitate retrograde, non-viral gene delivery to the liver via the bile duct

AIMS

To establish the extent of retrograde bile duct infusion at an ultrastructural level, as a preliminary step before evaluating the efficacy of gene delivery to the rat liver via a branch of the bile duct.

METHODS

The extent of retrograde infusion into the biliary tree was established by light and electron microscopy, following infusion of 10 nm gold particles into the right lateral lobe. Canalicular permeability was further assessed by the infusion of a 67 kDa protein. For gene delivery, both naked DNA and a synthetic peptide vector system were evaluated. Because canalicular tight junction permeability can be compromised in damaged livers, both normal rats and rats recovering from the hepatotoxin D-galactosamine were studied.

RESULTS

The gold particles penetrated the peripheral one third of the hepatic lobules and, surprisingly, reached the space of Disse in normal rats. Equally surprisingly, blood levels of a 67 kDa protein were identical after bile duct infusion and portal vein injection. Gene delivery with peptide/DNA complexes was much more effective in rats treated with D-galactosamine. However, gene delivery with naked DNA was equally effective in normal and damaged livers. Localisation of gene expression showed a scattering of positive hepatocytes restricted to the right lateral lobe.

CONCLUSIONS

Retrograde infusion into the bile duct advances well into the hepatic lobule and reveals a remarkable permeability of the canalicular or cholangiole tight junctions in normal rats. It is an effective approach for delivering genes to a small population (approximately 1%) of hepatocytes.

Is tolerance a prospective for clinical research?

Tolerance is an emotive issue in transplantation. It is the promised land for which we all strive and which we all hope we shall live to see. In such circumstances, tolerance must always be a prospective for clinical research! The question is, therefore, better posed in a more optimistic fashion and with a small act of faith: do we, in 1991, have that crucial combination of basic scientific knowledge and creative imagination to make it possible?

In vivo gene delivery via portal vein and bile duct to individual lobes of the rat liver using a polylysine-based nonviral DNA vector in combination with chloroquine

The objective of this study was to evaluate a bifunctional synthetic peptide as a DNA vector for regional gene delivery to the rat liver by the portal vein and bile duct routes. The 31-amino-acid peptide (polylysine-molossin) comprises an amino-terminal chain of 16 lysines for electrostatic binding of DNA, and the 15 amino acid integrin-binding domain of the venom of the American pit viper, Crotalus molossus molossus. Initial in vitro evaluation demonstrated that polylysine-molossin/DNA complexes were much smaller (approximately 50-100nm versus 500-1300nm), more positively charged, and more stable in isotonic dextrose in comparisons with salt-containing solutions. However, polylysine-molossin/DNA complexes in any solution other than complete culture medium were ineffective for gene delivery in vitro. Vector localization studies demonstrated that both the portal vein and bile duct routes provided excellent access of polylysine-molossin/DNA complexes to the liver. However, complexes delivered by the portal vein were rapidly lost (<15 min) following re-establishment of the portal circulation, whereas complexes delivered by the bile duct persisted much longer. Polylysine-molossin/DNA complexes in various isotonic solutions were delivered to the right lateral lobes either by perfusion through a branch of the portal vein or by infusion into appropriate branches of the bile duct. Two or three hours before gene delivery, rats were given a single injection of chloroquine. We report that the polylysine-molossin vector is much more effective (>10-fold) when delivered by the bile duct route with all isotonic solutions evaluated, and that polylysine-molossin/DNA complexes in isotonic dextrose are much more effective (>10-fold) than complexes in salt-containing solutions.

A powerful cooperative interaction between a fusogenic peptide and lipofectamine for the enhancement of receptor-targeted, non-viral gene delivery via integrin receptors

BACKGROUND

Following receptor-mediated endocytosis, vector/DNA complexes require assistance to exit endocytic vesicles in order to avoid degradation in the lysosomes. Overcoming this barrier is a major challenge for the development of receptor-targeted, non-viral gene delivery.

METHODS

The fusogenic peptide of influenza virus haemagglutinin, lipofectamine and chloroquine were tested singly and in combination in various doses for promoting in vitro gene transfer by an integrin-targeted, non-viral DNA vector (polylysine-molossin).

RESULTS

The fusogenic peptide and lipofectamine both individually promoted integrin-targeted gene delivery. However, the combined use of these agents was particularly effective, even at concentrations where neither agent singly had any effect on promoting gene delivery by polylysine-molossin. This optimal combination was effective on several cell lines and primary cell cultures. On the HuH7 cell line, it was approximately five-fold more effective than optimal chloroquine concentrations for integrin-targeted gene delivery and four to five times more effective than commercially available polyethylenimine. With the beta-galactosidase reporter gene, 60-65% of HepG2 cells and 75-80% of HuH7 cells were positive. The surface charge of polylysine-molossin/DNA/lipofectamine/fusogenic peptide complexes was approximately the same as that of polylysine-molossin/DNA complexes. The size distribution of the complexes suggested that competitive binding of polylysine-molossin and lipofectamine to DNA influenced the overall efficacy of this approach.

CONCLUSIONS

Although the mechanisms are not clear, the combined use of very low doses of two membrane-destabilizing agents results in high levels of receptor-targeted gene delivery.

The allogeneic response and tumor immunity

The strong allogeneic response to donor MHC molecules in transplantation and the weak response to tumor antigens represent two important and divergent but potentially interactive immune responses. A patient's response to allogeneic MHC molecules might promote an effective T-cell response to self MHC-restricted tumor peptides and the possibilities for this are discussed here. These allogeneic responses might successfully be harnessed to promote the immune eradication of metastatic cancer.

A small, synthetic peptide for gene delivery via the serpin-enzyme complex receptor

BACKGROUND

The serpin-enzyme complex receptor (SECR) has previously been successfully targeted for gene delivery using synthetic peptide ligands covalently linked in fluid phase to commercially available polylysine preparations (approximately 10-54kDa). The objective of the present study was to improve this approach by the use of small, bifunctional, and easily standardised synthetic peptides.

METHODS

Two synthetic peptides designated polylysine antitrypsin 1 (PAT1) (K16 FNKPFVFLI) and PAT2 (K16 CSIPPEVKFNKPFVFLI) were evaluated for gene delivery to the HUH7 human hepatocyte cell line. The K16 moiety binds DNA electrostatically, while the FVFLM motif of human alpha1-antitrypsin targets the SECR.

RESULTS

Both PAT1 and PAT2 bind to and condense DNA into small particles as shown by laser scattering techniques. However, only PAT2 is effective for gene delivery, presumably on account of the greater distance between the K16 chain and the FVFLM motif. Gene delivery by PAT2/DNA complexes is chloroquine-dependent, can be blocked completely by free ligand (CSIPPEVKFNKPFVFLI), and is highly efficient (e.g. approximately five-fold more effective than lipofectamine). At physiological salt concentrations, PAT2/DNA complexes formed at 4 microg/ml DNA are approximately 350 nm in diameter and highly effective for gene transfer, but at 100 microg/ml DNA the complexes are aggregated (diameter > 4 microm) and inactive.

CONCLUSIONS

A small (33 amino acid), bifunctional, synthetic peptide represents a highly efficient and readily standardised DNA vector for the SECR. The effectiveness of this peptide depends on the distance of the K16 moiety from the targeting ligand. High salt concentrations are not required to form effective vector/DNA complexes.

Efficient gene delivery to vascular smooth muscle cells using a nontoxic, synthetic peptide vector system targeted to membrane integrins: a first step toward the gene therapy of chronic rejection

BACKGROUND

Chronic rejection is now the major cause of allograft failure. A prominent characteristic of the histopathology is extensive intimal proliferation of vascular smooth muscle cells. Targeting vascular smooth muscle cells by gene therapy techniques offers a possible avenue for arresting or reversing chronic rejection. Defining suitable non-viral DNA vectors for this application is the objective of this study.

METHODS

A 31 amino acid synthetic peptide has been evaluated as a DNA vector for primary cultures of vascular smooth muscle cells of man, rabbit, and rat. The vector comprises a 15 amino acid integrin-binding domain and a chain of 16 lysines for electrostatic binding of DNA. Three agents known to promote exit of vector/DNA complexes from endocytic vesicles were studied systematically to define optimal, non-toxic conditions for gene delivery.

RESULTS

Initial binding studies on frozen sections showed that the integrin-binding domain binds strongly to vascular smooth muscle cells in all three species, thereby establishing vascular smooth muscle cells as a potential target for this receptor-targeted DNA vector system. Primary cultures of vascular smooth muscle were therefore studied. The use of chloroquine to assist endocytic exit, which works well on immortalized cell lines, was of little value because of toxicity to the primary vascular smooth muscle cells. The addition of cationic lipids to polylysine-molossin/DNA conjugates gave excellent reporter gene expression, but required mildly toxic doses of cationic lipid, and resulted in some loss of integrin specificity of the vector system. The optimal system involved the use of the amino terminal 20 amino acids of the hemagglutinin of the influenza virus. This peptide, when added to polylysine-molossin/DNA complexes at an optimal w/w ratio of 5:1:2 (polylysine-molossin/DNA/fusogenic peptide) resulted in 25-30% transfection of vascular smooth muscle cells with good levels of gene expression and no toxicity.

CONCLUSION

This represents an effective and safe DNA vector, comprised entirely of small synthetic peptides, and therefore readily standardized for clinical and experimental application.

An epithelial cell line that can stimulate alloproliferation of resting CD4+ T cells, but not after IFN-gamma stimulation

It has previously been shown that IFN-gamma-induced up-regulation of HLA class II on the surface of epithelial cells is not sufficient to induce proliferation of allospecific CD4+ T cells in vitro. To further investigate this phenomenon, a human epithelial bladder carcinoma, T24, was induced to constitutively express HLA class II without IFN-gamma stimulation, by permanent transfection with the full-length class II transactivator (CIITA) gene. Proliferation of allospecific T cells to transfected and wild-type cells with and without prior activation with saturating levels of IFN-gamma for 4 days was examined. IFN-gamma-activated T24 did not induce any response from CD4+ T cells. However, T24.CIITA induced significant levels of alloproliferation, which could be abrogated by pretreatment of T24.CIITA with a mAb to LFA-3. Prestimulation of T24. CIITA with saturating levels of IFN-gamma for 4 days also prevented allospecific CD4+ T cell proliferation. These findings suggest that epithelial cells may be intrinsically able to process and present alloantigen and provide adequate costimulation. We propose that IFN-gamma has a secondary, as yet unidentified, effect that acts to negatively regulate this response, at least in some epithelial cells.

Specific suppression of interleukin 2 biosynthesis by synthetic antisense oligodeoxynucleotides does not influence allograft rejection

BACKGROUND

Interleukin (IL)-2 supplementation can reverse both blood transfusion-induced tolerance to kidney allografts and spontaneous tolerance to liver allografts in rats. Moreover, IL-2 expression is frequently suppressed in models of allograft tolerance. The failure of IL-2 biosynthesis might therefore play a critical role in tolerance induction.

METHODS

Three antisense oligodeoxynucleotides (AS-1, AS-2, AS-3) to rat IL-2, and a control oligo (C-1) consisting of a scrambled version of AS-1, were evaluated for gene-specific suppression of IL-2 biosynthesis in vitro and in vivo, and for their effects on kidney allograft survival. Reverse transcriptase-polymerase chain reaction and IL-2 protein assays were used to assay concanavalin A-driven IL-2 biosynthesis by lymph node lymphocytes in vitro. PVG recipients of Dark Agouti kidney allografts were treated with the oligos. Graft survival and IL-2 biosynthesis by reverse transcriptase-polymerase chain reaction in spleen and graft biopsy specimens were assessed.

RESULTS

The AS-1 oligo, but not the AS-2, AS-3 or C-1 oligos, suppressed concanavalin A-driven IL-2 biosynthesis for the 4 days of culture. This effect was dependent on delivery of the AS-1 oligo with lipofectamine. Supplementation with exogenous IL-2 reversed the suppression of lymphocyte proliferation in AS-1-treated cultures. Administration of AS-1 intravenously at 10 mg/kg/day to PVG recipients of Dark Agouti kidney allografts suppressed IL-2 (but not IL-6, interferon-gamma, or tumor necrosis factor-alpha) synthesis in the grafts of seven of nine rats, as measured in biopsy specimens taken at days 2-7. By contrast, all nine control grafts strongly expressed IL-2. However, neither graft histopathology nor graft survival was affected.

CONCLUSIONS

Antisense oligonucleotides can powerfully suppress IL-2 biosynthesis in vitro and in allograft recipients in vivo, but this does not affect kidney allograft rejection.

In vitro investigation of factors important for the delivery of an integrin-targeted nonviral DNA vector in organ transplantation

BACKGROUND

Polylysine-molossin is a 31 amino acid synthetic peptide that has previously been demonstrated to function as a DNA vector in vitro for cell lines and for the cornea. It incorporates the 15 amino acid integrin-binding domain of the venom of the American pit viper, Crotalus molossus molossus as the targeting moiety and a chain of 16 lysines as the DNA-binding moiety. The objective of this study was to evaluate several parameters of importance for in vivo applications.

METHODS

Binding and tissue distribution of the vector/DNA complexes were followed by a monoclonal antibody to the vector, or by the use of fluorescein-labeled DNA. Standard in vitro transfections were used to monitor effective gene transfer.

RESULTS

(1) Optimal DNA/vector concentration. Saturation of vector/DNA binding sites on the ECV304 cell line occurred at 6 microg/ml of DNA. The concentration of vector/DNA complexes required for optimal gene transfection was found to be 2-8 microg/ml of DNA, corresponding to the concentration needed for saturation binding. (2) Optimal target cell exposure time. Vector/ DNA complexes saturated target cell binding sites within 5 min of incubation. However, lengthy exposure times (>2-3 hr) to the transfection medium were essential for substantial gene transfer. This was a consequence of two complementary factors. First, it was important that target cells be exposed to vector/DNA complexes for approximately 1 hr at 37 degrees C. Saturation of target sites at 4 degrees C and then removal of the transfection medium was much less effective. Second, exposure to chloroquine for 8-10 hr after uptake of vector/DNA complexes was essential for optimal gene transfer. (3) Inhibitory effects of serum. Exposure of complexes to even 1% serum before transfection, markedly inhibited gene transfer. However, target cells previously saturated with vector/DNA complexes and then exposed to 10% serum showed substantial gene transfer. (4) Extravasation and binding stability in vivo. Cold ex vivo perfusion of rat hearts with vector/DNA complexes demonstrated that little, if any, complex moved out of the vascular system. After transplantation of the heart, most of the complex bound to the vasculature was lost within 30 min of reestablishing the blood circulation.

CONCLUSIONS

Careful attention to several parameters of little importance in vitro need to be paid for optimal in vivo application of DNA vector systems.

Tissue-binding properties of a synthetic peptide DNA vector targeted to cell membrane integrins: a possible universal nonviral vector for organ and tissue transplantation

BACKGROUND

Gene delivery through a nonviral, receptor-mediated system widely expressed in transplanted tissue would have important advantages in transplantation, where gene delivery is performed ex vivo. Integrins are widely expressed cell surface receptors and can be targeted for gene delivery.

METHODS

A synthetic 31 amino acid DNA vector (polylysine-molossin) comprising a 15-amino acid moiety for targeting cellular integrins (derived from the snake venom, molossin) and a 16-amino acid polylysine moiety for DNA-binding, has been evaluated. The 31-amino acid vector, as well as its separate 15-amino acid integrin-binding and (lys)16 components, were individually synthesized, and a monoclonal antibody was raised to the molossin peptide for these studies. Binding to cell lines and tissue sections and capacity for gene delivery were examined.

RESULTS

Flow cytometric studies with the ECV304 cell line demonstrated that the binding of polylysine-molossin and polylysine-molossin/DNA complexes involved both electrostatic and integrin-mediated interactions with the cells, with the electrostatic binding being sufficient for maximal binding. However, binding to cellular integrins was essential for successful gene transfer. Binding studies on frozen tissue sections of the rat and pig demonstrated that the molossin peptide bound to many cell types of interest in transplantation, but not to all. Among the negative tissues were vascular endothelium and pancreatic islets. Small species differences in tissue binding were noted between the rat and pig.

CONCLUSIONS

This study defines the cooperative nature of the binding of this vector system to target cells and establishes the cell types most likely to be effectively targeted for DNA transfer.

The induction of major histocompatibility complex class II expression is sufficient for the direct activation of human CD4+ T cells by porcine vascular endothelial cells

BACKGROUND

The role played by major histocompatibility complex (MHC) class II-positive vascular endothelial cells in organ graft rejection is unknown but potentially very important. Methods. The MHC class II-negative porcine vascular endothelial cell line PIEC was stably transfected with the human class II transactivator CIITA, in order to induce MHC class II expression without the coinduction of T-cell costimulatory ligands. These PIEC cells were compared with interferon gamma-treated PIEC cells for their capacity to stimulate the proliferation of pure human CD4+ T cells. Results. The CIITA-transfected PIECs were as effective as interferon y-treated PIECs for stimulating unprimed human CD4+ T cells, the peak response with the CIITA-transfected cells in fact occurring earlier (day 3 instead of day 5). Monoclonal antibodies to SLA-DR substantially inhibited the CD4+ T-cell responses in both cases. However, whereas the response to interferon gamma-treated PIEC was partially inhibited by CTLA4-Ig, that to CIITA-transfected PIEC was not. Conclusions. The strong stimulation of CD4+ T cells by the specific induction of MHC class II antigens demonstrates that PIEC cells constitutively express functionally effective levels of costimulatory ligands. This finding strengthens the case that vascular endothelial cells are professional antigen-presenting cells and that MHC class II-positive vascular endothelial cells might play a role in the rejection of organ allografts.

Suppression of human anti-porcine T-cell immune responses by major histocompatibility complex class II transactivator constructs lacking the amino terminal domain

BACKGROUND

The class II transactivator (CIITA) is a bi- or multifunctional domain protein that acts as a transcriptional activator and plays a critical role in the expression of MHC class II genes. We have previously demonstrated that a mutated form of the human CIITA gene, coding for a protein lacking the amino terminal 151 amino acids, acts as a potent dominant-negative suppressor of HLA class II expression. Porcine MHC class II antigens are potent stimulators of direct T-cell recognition by human CD4+ T cells and are, therefore, likely to play an important role in the rejection responses to transgenic pig donors in clinical xenotransplantation. We were, therefore, interested in examining mutated CIITA constructs for their effect on porcine MHC class II expression.

METHODS

Stable transfectants of the porcine vascular endothelial cell line PIEC with mutated CIITA constructs were tested for SLA-DR and SLA-DQ induction by recombinant porcine interferon-gamma. Transient transfectants of the porcine B-cell line L23 with the mutated CIITA constructs were tested for the suppression of constitutive SLA-DR and SLA-DQ expression. T-cell proliferation studies were performed using highly purified human CD4+ T cells.

RESULTS

In preliminary studies, we demonstrated that transfection of the PIEC line with full-length human CIITA constructs resulted in strong expression of SLA-DR and SLA-DQ antigens, thus establishing the cross-species effectiveness of human CIITA in the pig. The mutated human CIITA constructs were, therefore, tested in the pig. PIEC clones stably transfected with one of these constructs showed up to 99% suppression of SLA-DR and SLA-DQ antigen induction and marked suppression of SLA-DRA mRNA induction. Moreover, transient transfection of the porcine B-cell line L23 showed up to 90% suppression of constitutive SLA-DR and SLA-DQ antigen expression in 5-8 days. In functional studies, interferon-gamma-stimulated PIEC clones transfected with this mutated CIITA construct failed to stimulate purified human CD4+ T lymphocytes.

CONCLUSION

Mutated human CIITA constructs are potent suppressors of porcine MHC class II expression.

Integrin-mediated transfection with peptides containing arginine-glycine-aspartic acid domains

Two synthetic peptides comprising an RGD moiety for integrin binding and a polylysine moiety for DNA binding were tested for transfection efficiency under a variety of different conditions. Binding of target cells to the peptide was shown to be strongly dependent on cyclisation of the peptides via cysteine residues. Low (10 microM) concentrations of chloroquine, added to assist endocytic exit, unexpectedly reduced transfection efficiency in two of the cell lines tested. COS-7 and ECV304. However, transfection efficiency increased at higher chloroquine concentrations and exceeded that in the absence of chloroquine in the case of the COS-7 and A375M cell lines. With the ECV304 cell line, optimum transfection occurred in the absence of chloroquine. Transfection efficiency of the peptides was greatest at peptide:DNA ratios of 4:1 (w/w), which were calculated to generate complexes containing approximately 5000 peptide molecules per plasmid. This represented approximately a 6:1 ratio of positive to negative charges. Peptide 5 was shown to have a higher transfection efficiency under most conditions, possibly because of more efficient stabilisation of cyclisation by two cysteine-cysteine bonds.

Suppression of MHC class II expression by human class II trans-activator constructs lacking the N-terminal domain

The class II trans-activator (CIITA) is a bi- or multi-functional domain protein which plays a critical role in the expression of MHC class II genes. We report that removal of the N-terminal 151 amino acids, encompassing all of the acidic domain but leaving intact the proline/serine/threonine-rich domain, results in a mutant protein with potent suppressive properties for MHC class II expression. HeLa cells stably or transiently transfected with mutant CIITA constructs showed up to 99% suppression of MHC class II antigen induction by IFN-gamma and marked suppression of HLA-DRA mRNA expression. Transient transfection of a B lymphoma line resulted in up to 89% reduction of constitutive MHC class II expression within 5 days and suppression of HLA-DRA mRNA synthesis.

A nonviral vector system for efficient gene transfer to corneal endothelial cells via membrane integrins

BACKGROUND

Genetic manipulation of allografts to suppress their ability to induce rejection is a promising approach for controlling rejection responses. A key to this approach is the development of appropriate DNA vectors. We are developing nonviral DNA vector systems based on synthetic peptides containing an integrin-binding segment for cellular targeting and a polylysine segment for DNA binding.

METHODS

Two such peptides have been tested for their ability to deliver the beta-galactosidase reporter gene to the corneal endothelial cells of the rabbit, pig, and man. One peptide was derived from a phage display library, the other from the integrin-binding moiety of the toxin from the American pit viper, Crotalus molossus molossus. Corneas were cultured overnight and then exposed to the DNA/peptide vector under a variety of conditions involving different DNA concentrations, chloroquine concentrations, times of exposure, presence of serum, and presence of polyanion buffers. Expression of the beta-galactosidase gene was determined after 3 additional days in culture. Effects of the treatment on the viability of the endothelium were examined by confocal microscopy.

RESULTS

We report that approximately 30% of corneal endothelial cells can be transfected with our optimal protocol using the molossin-based vector. Transfection is dependent on the presence of chloroquine and is inhibited by polyanion buffers such as HEPES. Viability of the corneal endothelium was excellent, except if corneas were incubated at high concentrations of chloroquine (0.5 mM) for prolonged periods (24 hr).

CONCLUSIONS

Synthetic peptides containing both an integrin targeting and a DNA-binding moiety are promising as simple and highly versatile DNA vectors for use in corneal transplantation.

Discordant expression of major histocompatibility complex class II antigens and invariant chain in interstitial dendritic cells. Implications for self-tolerance and immunity

BACKGROUND

The invariant chain plays a crucial role in antigen presentation by influencing the expression and peptide loading of major histocompatibility complex (MHC) class II molecules. Therefore, coordinate expression of these molecules is important for antigen presentation.

METHODS

Immunohistological studies were performed on frozen sections of many rat tissues in order to examine expression of invariant chain and MHC class II antigens.

RESULTS

Although coordinately regulated in most tissues, the interstitial dendritic cell (and the renal tubular epithelial cell) was always negative for invariant chain, while strongly positive for MHC class II antigens. However, renal tubular epithelial cells strongly expressed invariant chain during kidney graft rejection.

CONCLUSIONS

The absence of invariant chain in interstitial dendritic cells is unexpected, in view of their presumed function as sentinel antigen-presenting cells in the connective tissues. This might have important implications for antigen presentation for tolerance and immunity.

Organ culture of arteries for experimental studies of vascular endothelium in situ

The objective of this study was to determine whether organ culture of arteries could be used as a more physiological model than endothelial cell culture for the study of vascular endothelium in vitro. Small pieces of artery from rat, pig, piglet and man were cultured in 24-well plates for up to seven or eight days to study the characteristics of the vascular endothelial cell layer during the first week of culture, in particular its integrity, viability and propensity for cell division. Using conventional and confocal microscopy, silver-stained endothelial cell boundaries were shown to be intact at all time points, up to and including day 7. However, occasional very small gaps between endothelial cells were seen with the scanning electron microscope under high power at day 7. Using the bromodeoxyuridine technique, no endothelial cell division was seen at day 4 in any species, except for the occasional endothelial cell in rat aorta. At day 7, pig, piglet and human arteries showed only very occasional dividing endothelial cells, but many endothelial cells had divided by day 7 in rat aorta. Viability of the endothelium was assessed using fluorochromes and examination of the endothelial layer en face using confocal microscopy. Viability was always excellent (> 95%) up to day 4. By day 7, occasional patches of dead cells could be seen, which were most obvious in rat aorta. This study demonstrates that endothelial cells can be studied in situ in organ culture with intact morphology, lack of cell division and excellent viability for a minimum of four days. For many research questions involving vascular endothelium--for example the pathophysiology of hyperacute rejection--short-term organ culture of vessels is likely to represent a more physiological model than endothelial cell culture.

Indirect T-cell allorecognition and the mechanisms of immunosuppression by allogeneic blood transfusions

LEW rats given twice weekly intravenous transfusions of DA blood for 10 weeks showed a strong antibody response to intact DA class I antigens at day 7 that declined to undetectable levels by week 6. The response remained undetectable for the remainder of the course, in spite of the repeated transfusions of DA blood. At week 6 during the blood transfusion course, the LEW rats were immunised with a DA class I peptide known to be recognised by LEW CD4+T cells in a LEW APC-dependent manner. This resulted in the prompt reappearance of a strong antibody response to intact DA class I antigen. However, in vitro T-cell proliferation responses to peptide 1 appeared to be partially suppressed by the blood transfusions. Immunisation of LEW rats with this peptide 4 weeks before commencement of the course of DA blood transfusions prevented the decline in antibody levels normally seen during the blood transfusion course. These data indicate that the multiple blood transfusions are able to induce, in non-sensitised recipients, a reversible suppression of the indirect T-helper response specific for allogeneic peptides in the blood transfusion. Although our protocol of twice weekly transfusions does not correspond to the clinical pattern of blood transfusions, our results raise the possibility that antigenic cross-reactivity at the level of small polymorphic MHC peptides between blood and organ donors might represent the immunological basis for the beneficial effects of random blood transfusions.

Attempts to demonstrate indirect T cell allorecognition of donor MHC peptides in transplant patients

Indirect T cell allorecognition has been shown to play an important role in the rejection of allografts in experimental animals. Although there has been much speculation as to its role in clinical transplantation, especially with regard to chronic rejection, indirect T cell allorecognition has been difficult to demonstrate in transplant patients. In this paper, we looked for in vitro T cell proliferation to synthetic peptides corresponding to donor HLA-A and HLA-B incompatible antigens. Twelve 15 amino acid peptides corresponding to the hypervariable regions of six of the most common HLA class I alleles in Caucasian populations (A1, A2, A3, B7, B8 and B44) were studied. Blood was taken from 12 adult patients following one or more episodes of acute kidney graft rejection, and from three pediatric patients undergoing chronic rejection of heart/lung transplants. The donor-recipient combinations were selected such that at least one of the six HLA antigens above was present in the donor and absent in the recipient. Peripheral blood mononuclear cells from these patients responded strongly in proliferation assays to phytohemagglutinin. However, none responded to the incompatible donor HLA peptides. Compartmentalization of responding T cells, the effects of immunosuppression, and assay sensitivity are discussed as possible explanations for the negative results.

Comparison of adenovirus gene transfer to vascular endothelial cells in cell culture, organ culture, and in vivo

A replication-defective adenovirus 5 vector carrying the beta-galactosidase reporter gene was tested for its efficiency for gene delivery to vascular endothelial cells in various situations. Both porcine and human primary vascular endothelial cell cultures were very efficiently infected (>90%) at adenovirus concentrations of 10(10) pfu/ml or higher. Cultured rat fibroblasts and keratinocytes were even more readily infected, with >90% infection with adenovirus titers of 10(8) pfu/ml or higher. However, nondividing vascular endothelium in situ was very poorly transduced. Pieces of aorta from adult pigs, sheep, rabbit and rat, and pieces of human umbilical artery and vein were studied in organ culture. These showed only occasional positive vascular endothelial cells when exposed to the adenovirus vector at concentrations up to 5x10(11) pfu/ml. Kidney perfusion studies in rats and pigs gave similar results. The only exception to the above findings was in very young (3-4 day old) piglets, which showed excellent (>90%) infection of vascular endothelium with the adenovirus vector at titers of 10(10) pfu/ml. Our data suggest that adenovirus vectors will not be of value for gene delivery to uninjured vascular endothelium in situ, and are therefore unsuited for ex vivo genetic manipulation of vascular endothelium in organs for transplantation.

T and B cell responsiveness to donor class I MHC molecules and peptides in long survivors with kidney allografts

LEW rats with long-surviving (> 100 days) (DA x LEW)F1 kidney allografts were generated by treating the recipients with cyclosporine for 14 days after grafting. All rats were monitored after transplantation for the development of antibodies to intact donor class I MHC molecules. Cyclosporine completely suppressed the early antibody response to intact DA class I MHC molecules in all 19 LEW rats. However, 17 of the 19 rats developed antibodies between four and six weeks after grafting-i.e., between two and four weeks after the cessation of cyclosporine therapy, and maintained high levels of antibody to the donor class I molecules in spite of the long-term presence of the allograft. The 2 rats that did not produce antibodies to donor class I MHC molecules, along with one of the 17 that did produce antibodies, were immunized with a synthetic peptide corresponding to a region of the DA class I MHC molecule known to be recognized by LEW CD4+ T cells via the indirect recognition pathway. All 3 long survivors developed self APC-dependent CD4+ T cell proliferation to the immunizing donor peptides, and strong antibody responses to these peptides. However, none of these long survivors suffered rejection episodes as a consequence of the peptide immunization. In one of the two long-surviving rats without antibodies to intact donor class I MHC molecules at the time of peptide priming, the peptide priming resulted in the prompt development of strong antibodies to intact donor class I molecules. However, the other of these 2 rats did not produce such antibodies after peptide priming. Thus in this model of kidney allograft tolerance, with long-term exposure of the recipient's immune system to donor antigens without evidence of rejection, none of the animals develops tolerance for the indirect T cell recognition of donor class I MHC antigens. In occasional animals, B cells potentially reactive to intact donor class I molecules are present and are adequately exposed to antigen but are quiescent because of the absence of T cell help, perhaps as a consequence of reversible T cell suppression or anergy. In other occasional animals, B cell nonreactivity (anergy or tolerance) to intact donor class I molecules appears to develop.

A three-cell cluster hypothesis for noncognate T-B collaboration via direct T cell recognition of allogeneic dendritic cells

In this article, we propose that T cell help for B cells can occur via an unusual three-cell cluster, with recipient CD4+ T helper cells interacting via direct allorecognition with donor dendritic cell class II MHC antigens, recipient B cells interacting with MHC class I (or any other) antigen on the donor dendritic cell surface, and noncognate (i.e., antigen nonspecific) T-B collaboration. In this noncognate pathway, antigen processing by B cells is not required and T cell help is potent because of the high precursor T cell frequency for direct recognition of allogeneic class II MHC molecules. The data supporting this hypothesis are: 1. LEW rat strain recipients of interstitial dendritic cell-free (DAxLEW)F1 kidney allografts were shown to have no detectable antibody to donor class I MHC antigens at day 7 after grafting. By contrast, LEW recipients of normal (DAxLEW)F1 kidneys had strong antibody responses. 2. Consistent wih important role for donor dendritic cells in the early antibody response to donor class I MHC antigens was the finding that it was dependent on donor class II MHC antigens. PVG recipients, previously immunized with pure DA RT1.B class II MHC antigens, had virtually no antibody response to the class I MHC antigens of DA kidney allografts. 3. We confirmed the low and high responder status of PVG and LEW rats, respectively, to DA class I antigens by studying antibody responses to pure DA class I antigens. However, PVG and LEW recipients of DA kidney allografts did not differ in their antibody response to the donor DA class I MHC antigens. This is consistent with this response not requiring the processing and presentation of DA class I antigen by PVG recipients. 4. LEW recipients of interstitial dendritic cell-free (DAxLEW)F1 kidney allografts did eventually develop a strong antibody response to DA class I antigens, but this was delayed by several weeks. That this delayed antibody response was probably mediated by conventional T-B collaboration and that T help was rate limiting in this situation, was demonstrated by immunizing LEW recipients with a DA class I peptide. This markedly accelerated the kinetics of the antibody response to the dendritic cell-free (DAxLEW)F1 kidneys.

Heart transplantation for dilated cardiomyopathy

Between 1988 and 1994, 23 patients underwent heart transplantation for dilated cardiomyopathy. The age of the 13 boys and 10 girls was from 8 months to 16 years (mean 7.1 years). Selection criteria included failure to thrive despite maximal antifailure treatment and/or intravenous inotrope dependence. The aetiology of cardiomyopathy was idiopathic (n = 13), congenital (n = 3), anthracycline induced (n = 4), Barth's syndrome (n = 1), and maternal systemic lupus erythematosus (n = 2). The waiting period of heart transplantation ranged from one day to 147 days (mean 22 days). Maintenance immunosuppression included cyclosporin, azathioprine, and prednisolone. Follow up after transplantation was from one month to 62 months (median 27 months) with a mean actuarial survival of 95% at one year and 87% at three years. Four patients developed coronary artery disease, one of whom died as a consequence 15 months after heart transplantation. Heart transplantation has emerged as an acceptable therapeutic option, at least in the short term, for patients with dilated cardiomyopathy.

Indirect T cell allorecognition of donor antigens contributes to the rejection of vascularized kidney allografts

This report demonstrates for the first time that indirect T cell allorecognition of donor antigens can contribute to the effector mechanism of rejection of vascularized organ allografts. LEW (RT1(1)) rats were primed for indirect T cell allorecognition of DA (RT1av1) classical class I MHC molecules by immunization with synthetic 22-24 amino acid peptides corresponding to the alpha-helices of the RT1-A class I molecule. These rats received (DA x LEW) F1 kidney grafts that had been depleted of donor interstitial dendritic cells to minimize the direct T cell allorecognition response to the graft. The peptide-immunized rats rejected their grafts more rapidly than did control immunized rats, in terms of both graft function and survival. Moreover, the kinetics of antibody production to intact donor class I molecules after kidney transplantation was much more rapid in the peptide-immunized rats, suggesting that T cell help is the rate-limiting factor for antibody production to donor antigens in this model. It was of interest that we could not detect an antibody response to donor peptides after kidney graft rejection.

Fine specificity of peptide determinants for indirect T cell recognition of class I MHC alloantigens

CD4+ T cells from the LEW (RT1l) rat strain are able to recognize a 24-amino acid peptide from the alpha-helical region of the alpha 1-domain of the DA (RT1avl) RT1.A class I MHC molecule. This response is known to play a role in the effector mechanisms of rejection of DA grafts by LEW recipients. In this study, we demonstrate that the WAG (RT1u) strain is also able to respond to this peptide, but that the PVG (RT1c) strain does not respond. Fine specificity studies using a nested set of 15 mers derived from the 24 mer indicate that the LEW strain recognizes multiple T cell epitopes spanning the length of the peptide, while the WAG strain response is limited to the N-terminal region. With regard to B cell immunity, both the LEW and WAG strains give strong antibody responses when immunized with the free peptide, while the antibody response in the PVG strain is weak. Interestingly, in all 3 strains, the antibodies appear to be directed at the N- and C-terminal regions of the peptide, and to be almost entirely dependent upon the presence of the N- and C-terminal amino acids. These studies are potentially important when considering the specificity of rejection responses, and most particularly when considering the specific suppression of rejection mediated by indirect T cell allorecognition.

Physical mapping of the rat MHC class II genes shows a high level of interspecies conservation

We report here a pulsed-field gel electrophoresis map of the rat major histocompatibility complex (MHC) class II region. Using probes for the recently discovered Tap-1 and Tap-2 genes and the different MHC class II genes, we found the gene order in the rat MHC (RT1) region to be RT1.H-Tap-1-Tap-2-Bb-Ba-Db-Da. Moreover, the distance between the Tap-1 and the RT1.Da genes is approximately 150 kb. This, together with recent mapping of the RT1 class II region, demonstrates an extensive colinearity in the MHC region of different species.

T-cell recognition of an allogeneic RT1-Dbu class II MHC peptide

The allo-antibody response of several rat strains to an unconjugated synthetic 20 amino acid peptide derived from the alpha helical region of the RT1-Du beta chain was tested. The LEW (RT1l) and WAG (RT1u) strains produced little or no antibody; the PVG (RT1c) and DA (RT1av1) strains produced moderate amounts of antibody; while the BN (RT1n) strain produced strong primary and secondary antibody responses. This suggested that the BN strain was able to process and present the RT1-Dbu peptide on its class II molecules. In vitro proliferation studies demonstrated that LEW T cells did not respond to the peptide, whereas BN T cells responded strongly, and that the response in the BN strain was found only in the CD4+ T-cell subset. However, immunisation of BN rats with the RT1-Dbu peptide failed to cause any acceleration of rejection of WAG skin or kidney grafts. Moreover, BN rats primed with WAG skin and kidney grafts did not produce T cells reactive to the RT1-Dbu synthetic peptide. This suggests that the T-cell response of the BN strain to the synthetic major histocompatibility complex peptide was not relevant to the indirect T-cell allo-recognition response to naturally processed RT1-Du beta chains.

Long-term results of mastoid cavities grafted with cultured epithelium prepared from autologus epidermal cells to prevent chronic otorrhea

Chronic otorrhea and recurrent infection from open mastoid cavities are common and troublesome clinical problems for which there is no very satisfactory treatment. The authors have previously described a simple procedure to solve this problem, using autologous cultured keratinocyte layers grafted onto the unepithelialized open mastoid cavities. All procedures are carried out on an outpatient basis without anesthesia, except for local anesthesia for the skin biopsy. Twenty-six patients with 28 "difficult" mastoids, in which otorrhea had been present from 2 to 32 years, have been grafted with a follow-up period varying from 10 to 18 months. Seventeen mastoid cavities became symptom-free as a result of this technique. There were 4 failures, and partial success was shown in 7 cavities as judged by both the patient and by clinical examination.

Comparative analysis of the N-glycans of rat, mouse and human Thy-1. Site-specific oligosaccharide patterns of neural Thy-1, a member of the immunoglobulin superfamily

Protein structure and tissue type are known to influence glycosylation of proteins. We have previously investigated the N-glycans at each of the three glycosylation sites of the cell surface glycoprotein Thy-1 when isolated from rat brain and thymocytes. Here we report a comparative analysis of the site-specific N-glycosylation patterns from rat (Asn 23, 74, 98), mouse (Asn 23, 75, 99) and human (Asn 23, 60, 100) neural Thy-1. Despite considerable differences in amino acid sequence, the results show a remarkable conservation of the pattern of N-glycans at corresponding sites between the three species, as judged by chromatographic comparisons and glycosidase susceptibility. This is particularly marked for sites at Asn 74/75 in rat/mouse and the equivalent site at 60 in human Thy-1, as well as for sites at Asn 98/99 and 100, respectively. The sites at Asn 23 in rat/mouse also contained almost identical glycosylation patterns, but at this site human Thy-1 showed significantly different glycosylation patterns. These site glycosylation patterns are discussed in relation to the likely accessibility of the oligosaccharides for processing. It is known that within a species, the glycosylation of Thy-1 is tissue specific; therefore, this degree of conservation of glycosylation of Thy-1 expressed in the same tissue in different species is all the more striking, given the known variation between species in the amino acid sequence of Thy-1. It is therefore proposed that neural cells have a particular requirement for specific surface carbohydrates and that the Thy-1 polypeptide serves as an appropriate carrier for these structures.

A new technique using cultured epithelial sheets for the management of epistaxis associated with hereditary haemorrhagic telangiectasia

A new technique for the treatment of severe epistaxis associated with hereditary haemorrhagic telangiectasia is described. The nasal septum and inferior turbinates, surgically denuded of respiratory epithelium, were grafted using autografts of cultured epithelial sheets derived from buccal epithelium. All patients upon whom this technique has been used have shown considerable lessening in the frequency and severity of their epistaxes although two patients received grafts on two occasions, in each case approximately three months apart. It is postulated that a nasal lining of stratified squamous epithelium is likely to be more resistant to trauma than the normal respiratory type, and this is supported by the observation that bleeds very seldom occur from the oral cavity in this syndrome.

Indirect T cell allorecognition: a cyclosporin A resistant pathway for T cell help for antibody production to donor MHC antigens

LEW (RT1(1)) rats were primed for indirect allorecognition of DA (RT1avl) MHC molecules by immunizing either with synthetic peptides corresponding to the polymorphic regions of the RT1.Aavl classical class I MHC molecule, or with the isolated, denatured chains of the RT1.A class I, RT1.B alpha class II and RT1.B beta class II MHC molecules of the DA strain. These primed LEW rats received DA kidney allografts and were treated after grafting with cyclosporin A. Unprimed LEW controls mount a vigorous rejection response to DA kidney allografts and produce a strong antibody response to DA class I MHC antigens. Both the rejection and the antibody responses are virtually completely suppressed by cyclosporin A treatment in these controls. Priming to indirect allorecognition of donor MHC antigens did not diminish the effectiveness of cyclosporin A in suppressing the acute rejection of DA kidney grafts, but cyclosporin A could not suppress the early antibody response to the grafts in the primed rats. This finding could be of importance in clinical transplantation, where antibody-mediated graft damage might play an important role in both acute vascular rejection and chronic rejection.

T cell recognition of donor major histocompatibility complex class I peptides during allograft rejection

LEW (RTI1) recipients of DA (RTIav1) skin and kidney allografts were tested for the capacity of their T lymphocytes to proliferate to three 22-24-amino acid peptides from the hypervariable regions of the RTI-Aav1 classical MHC class I molecule. Ten days after rejecting second-set DA kidney allografts, spleen cells (but interestingly not lymph node cells) from LEW recipients showed strong, LEW antigen-presenting cell-dependent, CD4+ T cell proliferation to peptide 1 (from the alpha helical region of the alpha 1 domain). CD8+ T cells showed no response to peptide 1. There was no response by the spleen cells to peptide 2 (from the beta sheet of the alpha 2 domain) or peptide 3 (from the alpha helical region of the alpha 2 domain). Immunization of LEW rats with pure RTI-Aav1 class I H chain in Freund's adjuvant gave responses identical to that seen after grafting, i.e. good CD4+ T cell proliferation to peptide 1, but none to peptides 2 and 3. However, immunization of LEW rats with peptides 1, 2 and 3 in Freund's adjuvant resulted in good CD4+ T cell proliferation responses to each of the peptides. These data demonstrate that indirect allorecognition can be stimulated by allograft rejection, and emphasize that the physiological processing of donor antigens will influence which peptides will be important in indirect allorecognition in transplantation.

Rejection of skin allografts by indirect allorecognition of donor class I major histocompatibility complex peptides

LEW (RT1l) rats were immunized with peptides corresponding to the alpha helical region of the alpha 1 domain (peptide 1), the beta sheet of the alpha 2 domain (peptide 2), and the alpha helical region of the alpha 2 domain (peptide 3) of the RT1-Aav1 classical class I molecule of the DA (RT1av1) strain. The immunizations were without carriers, and the objective was to prime to indirect allorecognition without influencing direct recognition of the RT1-Aav1 molecule. The LEW rats mounted strong primary and secondary antibody responses to peptides 1 and 3, but only weak secondary responses to peptide 2. None of the antipeptide antibodies crossreacted with intact RT1-Aav1 class I molecules. The immunization also resulted in LEW antigen-presenting cell-dependent, CD4+ T cell proliferative responses, which were very strong against peptide 1 and weakest against peptide 2. LEW rats immunized with peptides 1 or 3, but most effectively with both peptides 1 and 3 together, showed accelerated rejection of DA skin allografts. This effect was not observed in LEW rats immunized with peptide 2. In response to the DA skin allograft, the peptide-immunized LEW rats showed markedly accelerated kinetics of antibody production to the intact RT1-Aav1 molecule. These data demonstrate that indirect allorecognition can play an important role in allograft rejection and have important implications for understanding allograft rejection and its regulation.

Stimulation of CD4+ T lymphocytes by allogeneic MHC peptides presented on autologous antigen-presenting cells. Evidence of the indirect pathway of allorecognition in some strain combinations

A preliminary analysis of the alloantibody response to free, unconjugated class I and class II MHC peptides in several rat and mouse strains was performed, to screen for an effective interaction between the allogeneic MHC peptides and recipient MHC molecules. The PVG rat strain was noted to produce very strong, MHC-restricted, primary and secondary responses to a synthetic peptide derived from the alpha helical region of the alpha 2 domain of an RT1.C/E class I MHC molecule of the DA strain. In vitro proliferation studies demonstrated that CD4+ but not CD8+ T cells of the PVG strain responded in a recipient APC-dependent manner to the peptide, whereas the BN strain (which showed no antibody response to this peptide) gave no T cell proliferation. Immunization of PVG rats with the peptide did not influence the rejection of DA skin allografts. The relevance of these studies to the possible mechanisms of allograft rejection by an indirect pathway are discussed.

Reduction of graft-versus-host reactivity after small bowel transplantation: ex vivo treatment of intestinal allografts with an anti-T cell immunotoxin

A specific T lymphocyte immunotoxin was used to pre-treat small bowel grafts in an attempt to prevent graft-versus-host (GVH) reactivity and GVH disease in a rat transplant model. The immunotoxin used was a conjugate of the anti-CD5 MoAb MRC OX-19 with ricin A chain. The grafts were perfused ex vivo with a standard solution of immunotoxin followed by incubation at 4 degrees C for 1 h before transplantation. In a semi-allogeneic strain combination (parent to F1 hybrid offspring) graft treatment with immunotoxin led to a prolongation of recipient survival compared with groups receiving similar transplants without immunotoxin treatment. An additive effect on survival was observed when the host was treated with cyclosporin. The effect of immunotoxin was greater than that of mesenteric lymphadenectomy in increasing host survival. The effect of graft treatment with the immunotoxin on cellular migration from graft to host lymphoid tissues was assessed in fully allogeneic transplantation (PVG to DA). Host lymphoid tissues were subjected to immunohistochemical analysis using a MoAb specific for donor class I MHC antigens. Graft treatment with the immunotoxin led to a significant decrease in the number of graft cells found in host lymphoid tissues 7 days after transplantation. However, this effect was less marked than that achieved by graft mesenteric lymphadenectomy. With our current protocol graft treatment with a specific T cell immunotoxin can significantly reduce but not abolish GVH reactivity in rat small bowel transplantation.