T-cell reconstitution without T-cell immunopathology in two models of T-cell-mediated tissue destruction

Abstract

Antigen-specific T cells play a pivotal role in adaptive immune responses. However, they also contribute to the progression of a variety of diseases including autoimmune disorders, graft rejection and graft-versus-host disease (GVHD). Non-specific immune-ablation treatments compromise the ability of the host to respond to infection, whereas the selective removal of epitope-specific T cells could theoretically ameliorate T-cell-mediated pathology while preserving the rest of the host immune function. In this study we investigated whether it is possible to destroy specific unwanted antigen-specific T cells by incubating polyclonal T-cell populations with major histocompatibility complex (MHC) tetramers that are conjugated to the ribosomal-inactivating toxin, saporin. This strategy resulted in a dramatic reduction in the number of targeted antigen (Ag)-specific CD8 T cells with no observable bystander toxicity in vitro. Moreover, in a model of transferable T-cell-dependent neurological disease induced by intracerebral (i.c.) lymphocytic choriomeningitis virus (LCMV) infection, the targeted killing of LCMV-specific CD8 T cells extended the survival of mice or fully prevented their death, depending on the dose of cells transferred. In addition, the tetramer– saporin conjugate also reduced liver damage in a model of donor T-cell-mediated hepatic destruction. These data provide a proof of principle that MHC tetramers could be exploited for the elimination or clinical manipulation of T-cell responses by linking effector molecules (a toxin in this case) to MHC tetramers. Also, the results suggest that it may be feasible to remodel T-cell responses, especially in immunocompromised hosts who receive adoptive cell transfers with many potential alloreactive cells.

Introduction

Throughout millions of years of selection, evolution has enlisted a very sophisticated group of cellular soldiers in chordates, which utilize their surface receptors to discriminate between allies and enemies.¹ These cellular soldiers, T cells, are usually involved in pathogen clearance and in cancer immunosurveillance. Nevertheless, there are times when T cells sense the body as foreign and elicit immune-mediated pathology, such as autoimmunity, organ rejection, or graft-versus-host disease (GVHD).^2–4 A common method for ameliorating these conditions is to use non-specific immunosuppressive agents. However, these usually render the host unable to mount protective T-cell responses against innocuous microorganisms. Also, in the context of irradiated patients who receive bone marrow transplantation, non-specific ablation of T-cell responses within the graft could reduce the efficacy of the graft-versus-leukaemia effect, even though it may mitigate the onset of GVHD.^5–7 Therefore, specific depletion of T-cell responses represents an important area for investigation.

The T-cell receptor (TCR) is internalized upon binding to the major histocompatibility complex (MHC).^8–10 Some research groups have tried to dissect antigen-specific T-cell populations in vitro using radioactively-labeled ¹¹¹In, and ²²⁵Ac ‘suicide’ tetramers, but these techniques have resulted also in the non-specific ablation of other T-cell populations,¹¹ which may be a result of bystander radioactivity emission.

In the present study we report the specific killing of an antigen-specific T-cell population by using MHC class I tetramers bound to the ribosomal-inactivating protein, saporin. We first performed this chemical conjugation early in 2006 (P. Penaloza-MacMaster, unpublished data). Coincidentally, it was first reported by Hess et al.¹² that it is possible to link saporin to MHC tetramers in order to develop a highly specific T-cell killing molecule. In this study, however, we were able to extend these findings to delay or fully prevent T-cell-mediated disease by using these antigen-specific suicide tetramers, thus providing the first clinical utility of this technology.

Saporin is a plant enzyme of 30 000 molecular weight with an N-glycosidase activity that specifically depurinates ribosomal RNA (rRNA) 28S, irreversibly blocking protein synthesis,^13–15 and has been used as a potent immunotoxin when bound to antibodies that recognize a specific cell receptor.^16–18 MHC class I tetramers, which provide the specificity for the delivery of the toxin, are regularly used for visualization of antigen-specific T cells and are produced by coupling biotinylated MHC class I monomers with streptavidin bound to a fluore.^19,20 In this study we used the commercially available saporin–streptavidin conjugate (Advanced Targeting Systems, San Diego, CA) instead of the fluore–strepatavidin derivative in order to produce potent T-cell-specific immunotoxins.¹²

In a model of lethal T-cell-mediated meningitis induced by adoptive T-cell transfer into irradiated mice, we showed that it is possible to reconstitute the T-cell repertoire in these immunocompromised mice without inducing T-cell immunopathology. The system for preventing T-cell-mediated tissue destruction involves the killing of antigen-specific encephalopathogenic T cells by treatment of polyclonal T-cell populations with saporin tetramer prior to adoptive transfer. T cells of other specificities remain viable. We also used mice that express a viral antigen in their liver to show how depletion of viral (liver)-specific T cells could ameliorate T-cell damage, without compromising the viability of other T-cell populations. These studies may be relevant for preventing GVHD in immunocompromised patients who receive bone marrow transplants containing alloreactive host-specific T cells. Also, in the context of graft-versus-leukaemia, it may be possible to percolate leukaemia-specific T cells from immunopathogenic T cells, thus providing maximal treatment efficacy.

Materials and methods

Mice and treatment of cells

Recipient 6 week-old C57BL/6 mice received sublethal irradiation of 550 rads. The following day, the mice received adoptive transfer of a mixture of OT-I and P14 splenocytes. These cells were stained with phosphate-buffered saline (PBS), H-2D(b) gp33 tetramer–allophycocyanin (APC), or H-2D(b) gp33 tetramer–saporin for 30 min at 21°. Cells were washed twice with 1% fetal bovine serum (FBS) in PBS (no azide). The following day, mice were infected intracerebrally (i.c.) with 10⁴ plaque-forming units (PFU) of lymphocytic choriomeningitis virus (LCMV) Armstrong.

Donor P14 Thy1.1⁺ mice and OT-I CD45.1⁺ transgenic mice were 6–8 weeks old, and were from our inbred colony. C57BL/6 mice were purchased from Jackson Laboratory (Bar Harbor, ME).

ALB1 transgenic gp33-expressing mice

This murine system for T-cell-mediated liver damage was a kind gift of Hanspeter Pircher. These mice received 10⁵ P14 cells and were then infected intraperitoneally (i.p.) with 2 × 10⁵ PFU of LCMV Armstrong.

Saporin tetramer preparation

Preparation of saporin-tetramer was carried out under similar conditions to that of standard MHC tetramer preparation. Saporin–streptavidin conjugate was obtained from Advanced Targeting Systems, and was diluted to 1 mg/ml. Ten reactions on ice were performed; each reaction lasted for at least for 10 min and followed the supplier’s precautions for the handling of saporin–streptavidin. Once the saporin suicide MHC tetramer was prepared, it was stored at 4°. Treatment conditions were the same as for staining of conventional tetramer, these tetramer staining recommendations are part of our lab's (and many others) standardized protocols. See Baylor College of Medicine website for a formal protocol: http://www.bcm.edu/proteincore/mhcstaining.html (1:100 for 10⁶ cells), mentioned elsewhere.

Serum alanine aminotransferase activity

Blood from the retro-orbital vein was collected and centrifuged at 3300 g for 30 min for separation of serum. Serum samples thus obtained were transferred to the Athens Diagnostic Laboratory (University of Georgia, Athens, GA) for analysis of serum alanine aminotransferase (sALT) activity (i.e. to measure the level of T-cell-mediated liver damage).

Results

MHC class I tetramers bound to the ribosomal toxin saporin are highly specific for their cognate T-cell population

As tetramers are internalized upon binding to their cognate TCR, staining of a mixture of OT-I cells [ovalbumin (OVA) specific)] and P14 cells (LCMV gp33-41 specific) with H-2D(b) gp33-41 tetramer linked to saporin would specifically kill only P14 T cells, while completely sparing OT-I T cells. In order to test this rationale, we stained ex vivo a mixture of P14 and OT-I cells with PBS, H-2D(b) gp33-41 tetramer–APC or H-2D(b) gp33-41 tetramer–saporin for 30 min at 21° (whole splenocytes were stained, and around 10% of cells were antigen-specific for the aforementioned epitopes). The cells were washed twice in cold PBS 2% FBS (no azide added) and then injected intravenously (i.v.) into C57BL/6 mice. Then, mice were bled on day 6 to determine the specificity of the surviving donor cell population after adoptive transfer of treated splenocytes (Fig. 1). As shown in Fig. 1, 6 days post-transfer, there was a markedly reduced number of gp33-41-specific T cells in the H-2D(b) gp33-41 tetramer–saporin group, whereas the non-targeted OT-I-specific population remained essentially unchanged. These findings were highly concordant with those of Hess et al.¹²

Figure 1.

Specific killing of antigen-specific P14 T cells without bystander cell depletion. A mixture of OT-I splenocytes (CD45.1⁺) and P14 splenocytes (Thy1.1⁺) was treated with phosphate-buffered saline treatment (tx) (PBS tx), H-2D(b) gp33-41 tetramer–allophycocyanin (APC), or H-2D(b) gp33-41 tetramer–saporin. Cells were washed, and 10⁶ transgenic (tg) T cells were injected intravenously (i.v.) into Thy1.2⁺CD452⁺ C57BL6/J recipients. Mice were bled on day 6 post-transfer, and peripheral blood mononuclear cells (PBMCs) were immunophenotyped for the presence of both donor populations. (a) P14 cells and (b) OT-I T cells. PBMCs were gated on CD3⁺ CD8⁺ lymphocytes. Representative fluorescence-activated cell sorter (FACS) plots from three separate experiments are shown (n = 2 mice per group).

Prevention of lethal T-cell-mediated meningitis by saporin tetramer: T-cell reconstitution without T-cell immunopathology

H-2(b) gp33-41 tetramer–saporin was found to be highly specific for its target P14 T cells. Therefore, we decided to develop a biological read-out to explore if we could prevent either morbidity or mortality after transferring antigen-specific T cells that would normally mediate severe tissue destruction. We took advantage of the LCMV murine i.c. infection model, which induces neurological signs of disease by day 5, including a visibly hunched posture, blepharitis and relative immobility of the mouse.^21,22 By day 8 the mouse has died of fatal convulsions, and histological analyses show severe destruction to the meninges, where LCMV antigens are found.^21,23 Development of lethal meningitis is T-cell mediated, as ablation of T cells by cyclophosphamide or other non-specific depletion therapy prevents disease.²⁴ The involvement of the T-cell response has also been verified using perforin knockout mice, in which i.c. infection with LCMV shows no observable lethality.²⁵

We recreated the models used by other groups.^24,26–28 This classical model of lethal choriomeningitis involves i.c. infection with LCMV, and death of the immunocompetent host by day 8 post-infection (Fig. 2a). However, immunosuppression results in the survival of mice. This is reversed by the adoptive transfer of immune splenocytes (Fig. 2b).

Figure 2.

Model for transferable, T-cell-mediated meningeal destruction. (a) and (b) Classical model for intracerebral (i.c.) lymphocytic choriomeningitis virus (LCMV)-specific induced lethality elucidated by Allan, Doherty, and others^22,24. (c) Our model involves the manipulation and dissection of antigen-specific T-cell responses before adoptive transfer, followed by a simple biological read-out to assess antigen-specific responses mediating immunopathology. Within each treatment group, the time of death of mice is usually directly correlated to the number of gp33-specific cells that are injected. APC, allophycocyanin; PBS, phosphate-buffered saline.

We tested the effect of antigen-specific T cells in the development of immunopathology by irradiating mice sublethally and by transferring polyclonal populations of T cells that were treated with PBS, standard MHC tetramer, or suicide tetramer. Then, the mice were infected i.c. with LCMV and observed for behavioral signs of meningitis and eventual death (Fig. 2c).

Adoptive transfer of OT-I cells does not mediate LCMV-induced lethality in irradiated mice infected i.c. with LCMV (P. Penaloza-MacMaster, unpublished data), whereas P14 T cells alone can do so if given in sufficient numbers. We observed lethal LCMV-mediated immunopathology following the adoptive transfer of splenocytes containing only 100 antigen-specific P14 T cells.

To model how depletion of immunopathogenic T cells from donor infusions could mitigate mortality and morbidity, we administered a sublethal irradiation of 550 rads to C57BL/6 mice. This was followed by adoptive transfer of a lethal dose of P14 T cells that were treated with PBS, H-2D(b) gp33-41 tetramer–APC or H-2D(b) gp33-41 tetramer–saporin (Fig. 3a). The following day, the mice were infected i.c. with LCMV Armstrong. Mice were observed for pathological signs of meningitis, and the time of death was recorded.

Figure 3.

Prevention or delay of T-cell-mediated lethality by treatment with H-2D(b) gp33 saporin tetramer. Mice were irradiated with 550 rads to remove all endogenous lymphocytes. Then, their haematopoietic repertoire was reconstituted with splenocytes containing P14 T cells treated with phosphate-buffered saline (PBS), H-2D(b) gp33-41 tetramer–allophycocyanin (APC), or H-2D(b) gp33-41 tetramer–saporin. Mice were then infected intracerebrally (i.c.) with 10⁴ plaque-forming units (PFU) of lymphocytic choriomeningitis virus (LCMV) Armstrong, and percentage survival was measured. (a) 10², (b) 10³, (c) 10⁴ and (d) 10⁵ex vivo-treated P14 T cells. n = 3 mice per group and data were combined from three separate experiments; P-values from the chi-square test are shown.

As Fig. 3 shows, treatment with the saporin–tetramer conjugate completely prevented death when the targeted antigen-specific population in the cell mixture was less than 10⁵ cells (Fig. 3a-c). However, when the number of targeted antigen-specific cells was equal to or higher than 10⁵, the H-2D(b) gp33-41 tetramer–saporin was only able to delay the onset of morbidity and mortality (Fig. 3d). Therefore, the depletion is not complete, and we believe that this method could be optimized or combined with additional treatment methods to induce a more specific depletion.²⁹ Notably, the mice that survived the first 15 days after i.c. infection usually did not succumb to lethal choriomeningitis, perhaps as a result of the establishment of regulatory mechanisms.

Interestingly, the mice that received P14 T cells treated with H-2D(b) gp33-41 tetramer–APC (normally used to detect antigen-specific T cells) died about 12 hr faster than those receiving PBS-treated cells, perhaps because of T-cell activation or tissue mobilization caused by the conventional MHC tetramer.

Amelioration of T-cell-mediated hepatitis by depletion of liver-specific T cells from donor polyclonal T-cell infusions

We then used the ALB gp33 mice, which express the immunodominant gp33 epitope of LCMV under the control of the albumin promoter.³⁰ The adoptive transfer of 10⁵ P14 cells, followed by infection with LCMV, induces transient hepatitis, as evidenced by an increase in sALT activity.³⁰ We adoptively transferred a mixture of OT-I and P14 T cells (10⁵ of each) that were treated with PBS, H-2D(b) gp33-41 tetramer–APC, or H-2D(b) gp33-41 tetramer–saporin (experimental scheme shown in Fig. 4). Transfer of the cell infusions that were treated with H-2D(b) gp33-41 tetramer–saporin resulted in decreased liver damage, evidenced by decreased sALT activity (Fig. 5). Extensive correlation between sALT activity and liver damage has been reported by Hess et al.¹²

Figure 4.

Model for transferable T-cell-mediated liver destruction. Variation of Hanspeter Pircher’s model [mice that express gp33 in hepatocytes are injected with P14 T cells and infected intraperitoneally (i.p.) with lymphocytic choriomeningitis virus (LCMV)]. Transient elevation in serum alanine aminotransferase (sALT) activity is reduced in transgenic (tg) mice that received cell infusions treated with H-2D(b) gp33-41 tetramer–saporin.

Figure 5.

Reduced T-cell-mediated liver damage [as measured by serum alanine aminotransferase (sALT activity)] after treatment of donor cells with H-2D(b) gp33-41 tetramer–saporin. (a)ALB treatment (tx) mice that express gp33 of the lymphocytic choriomeningitis virus (LCMV) were given intravenous (i.v.) infusions of splenocytes containing 10⁵ P14 T cells treated with either phosphate-buffered saline (PBS) or H-2D(b) gp33-41 tetramer–saporin. Mice were then infected intraperitoneally (i.p.) with 2 × 10⁵ plaque-forming units (PFU) of LCMV Armstrong and the sALT activity was measured on days 0, 5, 10 and 20 post-infection. n = 2, data combined from two separate experiments. Data shown represent the average ± standard error of the mean (SEM).

A very small population of antigen-specific T cells that survive depleting treatment could expand extensively upon challenge, although less in absolute numbers compared with the PBS control

We then decided to analyze the kinetics of T-cell expansion of the very few surviving P14 cells that were not killed by the gp33 tetramer–saporin. We treated a mixture of Thy1.1⁺ P14 T cells and CD45.1⁺ OT-I T cells with PBS, H-2D(b) gp33-41 tetramer–APC, or H-2D(b) gp33-41 tetramer–saporin for 30 min at 21°. Cells were then washed and injected i.v. into Thy1.2 B6 mice. Mice were then challenged i.p. with LCMV at day 6·5 post-transfer in order to test for expansion of any unkilled donor cells.

As Fig. 6 shows, even though treatment with saporin tetramer resulted in a pronounced reduction of P14 T cells in peripheral blood mononuclear cells (PBMCs) by day 6·5, upon challenge with 2 × 10⁶ PFU of LCMV, a minute fraction of survivors remained in the H-2D(b) gp33-41 tetramer–saporin treated group that expanded robustly upon challenge. The reasons for this resistance to saporin may include relative metabolic quiescence (so transient ribosomal inactivation may not harm these cells), TCR down-modulation by some T cells (initially, they may appear invisible by tetramer staining), or simply that the treatment protocol may be amenable to optimization. However, as expected, the numbers of donor T cells following challenge were less than in the groups treated with PBS or APC tetramer. Interestingly, in the saporin tetramer-treated group, endogenous antigen-specific T cells (Thy1.1⁻) are evident, whereas the appearance of endogenous T cells in the undepleted groups (PBS, and H-2D(b) gp33-41 tetramer–APC) is precluded. This is probably a result of less competition between donor and endogenous gp33-specific T cells.³¹ Therefore, even if 99·9% depletion occurs, from 10⁵ to 10² P14 cells, these 10² P14 cells that survive the depletion could in theory reconstitute the original endogenous gp33 specific population present before irradiation. This original population number is estimated to be as small as 100–200 cells.³¹ This is corroborated by the fact that we observed lethality in the i.c. LCMV model only when we adoptively transferred 100 or more P14 T cells into irradiated mice, bringing the gp33-specific population to its original number.

Figure 6.

Very few P14 T cells surviving treatment with the tetramer–saporin conjugate could still expand robustly upon challenge. Naïve P14 Thy1.1⁺ splenocytes were treated with phosphate-buffered saline (PBS) treatment (tx), conventional tetramer [H-2D(b) gp33 allophycocyanin (APC) treatment (tx)], or tetramer gp33 saporin [H-2D(b) gp33 saporin tx] for a lapse of 30 min on ice. Cells were then washed and injected intravenously (i.v.) into Thy1.2⁺ B6 mice. Peripheral blood mononuclear cells (PBMCs) were collected on days 3, 6·5, 8 and 18, before and after infection with LCMV. Upon LCMV challenge, the mice that received tetramer–saporin -treated cells underwent more endogenous (Thy1.1⁻) proliferation of gp33-specific cells as a result of less competition for antigen. Undepleted OT-I T cells were also tracked throughout this kinetic experiment and their numbers were similar in all groups (some contraction is seen after P14 expansion). The PBMCs shown were gated on CD3⁺ CD8⁺ lymphocytes.

Discussion

The removal of epitope-specific T cells may be especially important in GVHD, or in transfusion associated-GVHD assuming that unwanted T-cell responses are known. We have shown, in the T-cell-mediated choriomeningitis model, that it is possible to prevent lethality if donor cells of various specificities are treated with H-2D(b) gp33-41 tetramer–saporin. However, if a very high number of antigen-specific cells are transferred, our saporin tetramer is only able to slow down morbidity and mortality. Therefore, there may be room for optimization of this protocol. Another possibility could be that there may be T cells that down-modulate their TCR and may be invisible to the tetramer–saporin, but once they are transferred into the host they undergo extensive activation and division upon antigen challenge.

In vivo injection with tetramer–saporin to remove encephalopathogenic T-cell populations may also be feasible in this model.¹² In our studies, however, even though we observed some depletion of antigen-specific cells when tetramer–saporin was injected directly into the mice, we noticed variability within groups, and sometimes death, after in vivo administration of tetramer–saporin. Tetramer–saporin has been reported to cause transient hepatotoxicity,¹² so conjugation with milder toxins are expected to be necessary for safe clinical use.

We also noticed that the standard fluore-conjugated tetramers seem to alter trafficking of antigen-specific T cells when injected i.v. into P14 chimeric mice (P. Penaloza-MacMaster, unpublished data). Donor P14 cells (Thy1.1⁺) disappear from the blood within 30 min after tetramer injection. By contrast, donor P14 T cells in the spleen remain there after i.v. administration of tetramer (P. Penaloza-MacMaster, unpublished data).

The presumed mechanism of depletion mediated by saporin-conjugated tetramer involves the release of saporin toxin from the tetramer upon its internalization by the antigen-specific T cell. This technology to dissect antigen-specific T-cell responses may help us to understand the contribution of epitope-specific responses in pathogen clearance by depletion of one or more T-cell specificities in an immune host and observing how well the pathogen is cleared when one or more T-cell responses is collapsed. Knowledge of which specific T-cell responses are important for immune control could guide vaccine research. It is also of relevance that in some chronic viral diseases, such as human immunodeficiency virus (HIV), specific T cell responses, such as that for env-polyprotein, seem to be correlated with increased viraemia, whereas other T-cell responses such as that for gag polyprotein, are correlated with reduced viraemia.³² By assuming that env-specific T cells directly induce an increase in the viral load (a cause rather than a consequence), it would be interesting to observe if the depletion of env-polyprotein-specific T cells would result in a reduction of the viral load.

Modifications in the tetramer conjugation chemistry may open new avenues into T-cell therapy, involving perhaps the targeting of antigen-specific T cells with tetramers containing thiol-modified microRNAs (which would presumably dissociate from the MHC tetramer upon endocytosis), as a way to modulate specifically the expression of a T-cell protein or ribosomal activity. Of particular interest is miR-181a, a miRNA that is down-regulated on mature peripheral T cells and suppresses the expression of protein phosphatases.³³ Expression of this miRNA in T cells is directly correlated to T-cell sensitivity and function,^33,34 and we propose that chemically linking miR181a or any nucleic acid message to streptavidin via a disulfide bond (which would break upon internalization) could be an easy way to direct a new activation/sensitivity programme to antigen-specific T cells. At least, it has been shown that it is possible to deliver a death message specifically to particular T cells in a mixture of other T cells and to direct their fate without affecting the fate of other T-cell populations. It may be interesting to exploit the streptavidin–biotin chemistry to deliver messages to antigen-specific T cells. It would be of interest to tetramerize biotinylated MHC monomers with streptavidin–oligo conjugates and to test for the expression of that oligo in antigen-specific T cells (future goals). Also, in animal models, this technology may be helpful in analyzing the contribution of each antigen-specific response by adoptive transfer of immune splenocytes into Rag^−/− or irradiated hosts. Adoptively transferred cells would be previously depleted of particular T-cell populations, and clearance of a pathogen could be analyzed in the presence or absence or specific responses.

In the near future, the MHC tetramer technology may be also important for treating T-cell exhaustion, such as that observed in acquired immune-deficiency syndrome (AIDS) (perhaps by coupling MHC tetramers to oligos that code for desirable phenotypic or activation markers on viral-specific, but not on self-reactive, T cells). At least it is now evident that one can redirect the fate of an antigen-specific T cell without affecting other cells. Instead of delivering a death message, however, it may be possible to deliver a survival message to particular T-cell clones. Perhaps even autoimmunity could be treated using this technique, assuming that the unwanted T-cell populations are known.

Disclosures

The authors declare no competing interests.