Donor Lymphocyte Infusion–Mediated Graft-versus-Host Responses in a Preclinical Swine Model of Haploidentical Hematopoietic Cell Transplantation

Abstract

We previously described successful hematopoietic stem cell engraftment across MHC barriers in miniature swine without graft-versus-host disease (GVHD) using novel reduced-intensity conditioning regimens consisting of partial transient recipient T cell-depletion, thymic or low-dose total body irradiation, and a short course of cyclosporine A. Here we report that stable chimeric animals generated with these protocols are strongly resistant to donor leukocyte infusion (DLI)-mediated GVH effects. Of 33 total DLIs in tolerant chimeras at clinical doses, 21 failed to induce conversion to full donor hematopoietic chimerism or cause GVHD. We attempted to overcome this resistance to conversion through several mechanisms, including using sensitized donor lymphocytes, increasing the DLI dose, removing chimeric host peripheral blood cells through extensive recipient leukapheresis before DLI, and using fully mismatched lymphocytes. Despite our attempts, the resistance to conversion in our model was robust, and when conversion was achieved, it was associated with GVHD in most animals. Our studies suggest that delivery of unmodified hematopoietic stem cell doses under reduced-intensity conditioning can induce a potent, GVHD-free, immune tolerant state that is strongly resistant to DLI.

INTRODUCTION

In the treatment of hematologic malignancies with allogeneic hematopoietic cell transplantation (allo-HCT), the eradication of residual tumor cells by donor-derived T cells is known as the “graft-versus-leukemia” effect. This is one of the most potent mechanisms by which allo-HCT leads to the resolution of cancer [1,2]. Unfortunately, the beneficial effects are often complicated by graft-versus-host disease (GVHD) [3,4]. GVHD is often observed with aggressive myeloablative regimens, and therefore new protocols using minimal and transient immunosuppression are of great interest [5,6]. The goal of reduced-intensity conditioning (RIC) regimens is to harness the graft-versus-leukemia effect while minimizing the nonhematopoietic system effects, which manifest as GVHD. RIC regimens have provided an opportunity for older patients with significant comorbidities, who were previously seen as unfit for transplant, to receive an allo-HCT [6,7].

The use of donor leukocyte infusion (DLI) to treat leukemic relapse was reported by Kolb et al. in 1990 [8]. Since then, the use of DLIs after allo-HCT has expanded [9]. DLI after HCT has been used for several indications, including as a prophylactic therapy for patients with a high risk of relapse, treatment of post-transplant lymphoproliferative disease, treatment of viral infections, and graft failure. However, the most common indication remains the treatment of relapsed malignancy. The use of DLI has been very effective as a therapeutic option for chronic myeloid leukemia but less so for acute myeloid leukemia [9]. Similarly to allo-HCT, a major complication of DLI is GVHD. Several retrospective analyses of GVHD after DLI have reported an overall incidence of acute GVHD of 40% to 60%, with 20% to 35% reporting grades III to IV disease [9]. Studies have also demonstrated a correlation between graft-versus-leukemia and GVHD after DLI. In a multicenter study assessing 140 patients who received a DLI for relapsed malignancy after stem cell transplantation, 93% of patients who achieved a complete response developed acute GVHD, whereas only 13% of patients who did not develop acute GVHD achieved a complete response [9].

Previously, we reported a novel RIC allo-HCT protocol that consisted of partial, transient recipient T cell depletion, thymic or low-dose total body irradiation, and a short course of cyclosporine A (CyA) that resulted in stable mixed chimerism, donor unresponsiveness in vitro, and permanent hematopoietic stem cell engraftment across MHC barriers without GVHD [7,8]. Mixed chimerism has been shown to induce long-term immune tolerance through central deletion of donor-reactive thymocytes. After several unmodified DLIs at clinical doses did not successfully result in conversion to full donor chimerism, we hypothesized the presence of a regulatory cell population potentially involved in the resistance to conversion in this model. We attempted to overcome this resistance through several mechanisms, including the use of sensitized donor lymphocytes, increasing the DLI dose, removing chimeric host peripheral blood cells through extensive recipient leukapheresis immediately before DLI, and using fully mismatched lymphocytes. Despite our attempts, the resistance to conversion in our model was robust, and when conversion was achieved, it was associated with GVHD in most animals.

METHODS

Animals

Transplant donors and recipients were selected from our herd of partially inbred, MHC defined miniature swine, which have been selectively bred over the past 40 years for large animal studies of transplantation [10]. Through several generations of inbreeding, these swine have been defined at the MHC genes encoding for class I and class II antigens. This has allowed for reproducible transplantation studies across defined genetic haplotypes. In this study donors ranged from 4 to 8 months of age and recipients from 9 to 12 weeks of age at the time of HCT. Donors and recipients were chosen to differ by 1 MHC haplotype at both MHC-I and MHC-II, mimicking a parent to child haploidentical transplant. To facilitate the detection of chimerism after HCT, only donors who were positive for pig allelic antigen (PAA) were selected. PAA is a nonhistocompatibility cell-surface antigen that is present on all differentiated hematopoietic cells in animals that express this gene allele [11]. PAA-negative recipients were selected.

HCT Protocol

The HCT protocol consisted of a combination ofeither 1000 cGy thymic or 100 cGy total body irradiation, T cell depletion, and hematopoietic cell infusion from a single-haplotype MHC-mismatched donor, with 30 to 60 days of CyA forthe peri-/post-transplant period. Partial and transient host T cell depletion was achieved using a single dose of the chemically conjugated pCD3-CRM9 immunotoxin on day −2 or a recombinant CD3-immunotoxin administered twice daily from day −4 to day −1 [12,13].

Donor animal hematopoietic cells were mobilized with either a combination of recombinant porcine IL-3 and porcine stem cell factor(Immerge Biotherapeutics, Cambridge, MA), each at a dose of .1 mg/kg for the first 30 kg and .05 mg/kg for each additional kilogram, or by recombinant human granulocyte colony-stimulating factor at a dose of 10μg/kg (Filgrastim; Amgen, Thousand Oaks, CA) for 5 to 7 days before day 0 [14]. Peripheral blood mononuclear cells (PBMCs) were collected by leukapheresis (COBE BCT Inc., Lakewood, CO) beginning on the fifth day of cytokine therapy and continuing until the target cell number was attained. After the initial leukapheresis, 1 to 15 × 10⁹ PBMCs/kg were infused intravenously daily until the target dose was achieved. Enteral CyA was administered via a gastrostomy tube, beginning 1 day before mobilized PBMC infusion and continuing for 30 to 60 days. CyA whole blood levels were maintained between 400 and 800 ng/mL for the first 30 to 45 days before being tapered for the last 15 days to a level of 200 ng/mL, at which point CyA was discontinued.

Unmodified DLI

Nonmobilized leukocytes were collected by leukapheresis from either the original hematopoietic cell donor (swine leukocyte antigen SLA^ac, referred to as AC), an (AC) MHC matched donor, or a (swine leukocyte antigen SLA^cc, referred to as CC) mismatched donor. Lymphocytes were infused intravenously into the recipient at a normalized dose (CD3⁺ T cells) at doses from 1 to 150 million T cells/kg.

Sensitized DLI

Some donors were purposely sensitized before DLI collection by receiving a fresh (nonfrozen) split-thickness skin graft (4 × 6 cm) from an animal MHC matched (swine leukocyte antigen (SLA)^ad, referred to as AD) to the host chimera. Complete rejection of skin grafts was observed by day 14. Between days 90 and 100, cell-mediated lymphocytotoxicity (CML) assays were performed assessing donor (AC) antichimera responses and compared with a naive donor (AC) antichimera response within the same assay. Donor animals were then leukapheresed forthe harvest of “sensitized” lymphocytes, which were used for DLI.

Criteria of a DLI Response

DLIs were considered to be effective when any of the following (or multiple) changes were observed: (1) a sustained increase in the percent of donor cells in the peripheral blood after DLI, (2) bone marrow (BM) effects defined as BM failure or donor reconstitution, and (3) development of GVHD. The small rise in lymphocyte chimerism inevitably seen immediately after DLI infusion was not considered an effect of the DLI.

Assessment of GVHD

Animals were monitored for development of GVHD by daily clinical examination and blood counts. A GVHD scoring sheet that assesses skin lesions, diarrhea, appetite, and liver function tests was also used to “score” GVHD [15]. Skin, BM, thymus and rectal biopsies were taken before conditioning and at 7-week intervals after HCT. If GVHD was suspected, additional biopsies were taken for comparison.

Flow Cytometry

Flow cytometry was performed using a FACScan (Becton-Dickinson, San Jose, CA). The following swine specific antibodies were used: CD3ε (898H2-6-15, mouse IgGaK) [16], CD4 (74-12-4, Mouse IgG2bK), CD8α (76-2-11, mouse IgG2aK), CD172 (74-22-15, mouse IgG1K)[17–20], CD5 [21], CD1 (767-4, mouse IgG2aK), CD16 (G7, mouse Ig), PAA (1038H-10-9,IgMK), CD25 (231.3B2), and FoxP3 (FJK-16s). For assessment of chimerism, PAA staining was used to distinguish between cells of donor and recipient origin [11]. Monocyte and granulocyte chimerism was determined by gating on CD172-positive mononuclear cells and granulocytes, respectively.

Assessment of Chimerism

Peripheral blood, BM, and thymic chimerism were assessed by flow cytometry as described previously[11,14,22]. Detection of donor-derived BM colony-forming units longer than 14 weeks after HCT correlates with engraftment of hematopoietic stem cells [11,14,22,23]. Thymic chimerism and multilineage peripheral blood chimerism present at 14 weeks always correlate with the presence of BM colony-forming units [23]. Consequently, engraftment was defined as the presence of any of these 3 markers (BM colony-forming units, thymic chimerism, or multilineage peripheral blood chimerism) after week 14 post-HCT.

CML Assay

CML assays were performed as previously described [24,25]. Briefly, responders (4 × 10⁵ cells) and stimulators (4 × 10⁵ cells; irradiated with 25 Gy) were cultured together for 5 days. Effector cells (term used to describe responder cells after 5 days of co-culture) were then replated with chromium-51 labeled target cells at effector-to-target ratios of 100:1, 50:1, 25:1, and 12.5:1. PBMCs from non-inbred Yucatan or Yorkshire pigs were used as third-party positive control stimulators and targets. The percent specific lysis was calculated as previously reported [24–26]: percent specific lysis = average experimental release (cpms) – average spontaneous release (cpm)/average max release (cpm) – average spontaneous release (cpm).

STATISTICS

A repeated measures analysis of variance with Bonferroni correction for multiple comparisons was performed. We compiled the data of the anti-self, anti-donor, and anti-third party cytolytic responses at each ratio and assessed for significance <.05.

RESULTS

Swine Undergoing Haploidentical HCT with RIC Regimens Are Unresponsive to Donor and Resistant to DLI

Figure 1 shows a representative group of 3 tolerant mixed chimeras generated by our RIC-HCT protocol, as defined by successful engraftment and unresponsiveness to donor MHC by CML. In total, 33 delayed DLIs were delivered to 26 RIC-HCT chimeric recipients. The initial DLI doses ranged from 10 to 50 × 10⁶ CD3+ Tc/kg, which are based on the range of clinical doses considered in Massachusetts General Hospital clinical practices (T.R. Spitzer, personal communication). DLIs were delivered at least 30 days after HCT and ranged from day +35 to day +984 after HCT. Of the 33 total DLIs, 21 (64%) had no effect on chimerism and did not cause GVHD (Tables 1 and 3) and 12 (36%) did have an effect (Table). Twenty DLIs that had no effect are provided in Table 1. No recipients from Table 1 exhibited a sustained increase in donor peripheral blood chimerism, developed GVHD, or underwent BM failure or donor reconstitution. This demonstrates that most tolerant mixed chimeras transplanted with our protocol are resistant to conversion after DLI with doses up to 50 × 10⁶ CD3⁺ Tc/kg.

Figure 1.

Long-lasting donor unresponsiveness post-HCT. Three representative long-term tolerant animals that received a DLI after a RIC-HCT. Animal 13272 received 2 DLIs (arrows) on days 745 and 984 post-HCT, respectively. The first DLI was unmodified, whereas the second DLI was a “sensitized” DLI. Animal 14225 received a DLI on day 110 post-HCT. Animal 13476 received a DLI on day 482 post-HCT. Chimerism is shown in the top panels. Percent donor-derived chimerism is shown on the y axis and days post-transplantation on the x axis. CMLs were done at different time points. White squares represent anti-self responses, triangles anti-third party responses, and dark squares anti-donor responses. The y axis represents percent specific lysis (PSL) and the x axis represents the effector-to-target ratio (E:T). No effects after DLI were observed in these 3 representative animals (*P < .05).

Table 1.

Resistance to Conversion after DLI

Animal No.	Haplotype Mismatch	Type of DLI	PB Chimerism at Time of DLI	GVHD	DLI
13272	AC→AD	DMLI	L	N	Day +745
		S. DMLI	L		Day +984
13476	CD→AD	DMLI	L	N	Day +482
13810	AC→AD	DLI	L	N	Day +82
14225	AC→AD	DLI	M	N	Day +110
14375	AC→AD	DLI	M	N	Day +35
14376	AC→AD	DLI	M	N	Day +35
		S. DLI	L		Day +287
14529	AC→AA	DLI	M	N	Day +35
14547	AC→AD	DLI	M	N	Day +35
14548	AC→AD	DLI	M	N	Day +35
16558	AC→AD	DLI	M	N	Day +152
16626	AC→AD	DLI	M	N	Day +153
		DMLI	M		Day 566
17017	AC→AD	DLI	M	N	Day +145
17467	AC→AD	DLI	M	N	Day +62
17469	AC→AD	DLI	M	N	Day +62
18862	AC→AD	DLI	-	N	Day +175
19125	AC→AD	BMT/DLI	L	N	Day +60
19126	AC→AD	BMT/DLI	L	N	Day +60

Seventeen animals who underwent DLI or donor-matched leukocyte infusion (DMLI) and had no effect defined as either a sustained increase in donor chimerism, BM failure, or GVHD. Two animals (19125, 19126) received a BM Transplant (BMT) and DLI. Animal 18862 is a control animal that had regained antidonor responses and had lost chimerism. S indicates sensitized; N, no; Y, yes; Chimerism: L, lymphoid alone; M, multiple lineage (both lymphoid and myeloid lineages); -, no chimerism.

Table 3.

Leukapheresis before DLI

Animal No.	Day of Leukophoresis Post-HCT	Amount of PBMCs Leukophoresed	PBMC × 107/kg Removed	Effect	GVHD
17467 (22 kg)	96	1.2 × 109	5.45	Yes	NA
17469 (16 kg)	60	4.5 × 109	28	No	No
19560 (20.1 kg)	57	4.5 × 109	22.5	Yes	Yes
19561 (19.4 kg)	58	1.5 × 109	7.8	Yes	Yes

Four animals were leukaphoresed before DLI. Three of 4 had a GVH response. Animal 17468 had a mild increase in chimerism but died before full the effects of the DLI could be assessed. Animals 19560 and 19561 did exhibit an increase in peripheral blood chimerism. Animal 19560 died early from sepsis secondary to BM failure after the DLI (GVH reaction). Of note, this animal had recently recovered from post-transplant lymphoproliferative disease after CyA was discontinuation (rapid taper). Animal 19561 converted to full donor and developed GVHD. Animal 17469 had no increase in peripheral blood chimerism.

Of the 33 DLIs performed in tolerant mixed chimeras, 12 DLIs had an effect as summarized in Tables 2 and 3. Of these 12, only 1 animal (no. 14980) did not develop GVHD after DLI. Animal no. 14980 received a 50 × 10⁶ Tc/kg DLI on day +219 post-HCT (Figure 2) and converted to full donor chimerism.

Table 2.

GVH Response Post-DLI

Animal No.	Haplotype Mismatch	Type of DLI	Engrafted	PB Chimerism at DLI	GVHD?	Dose (cells/kg)
DLI effect (eg, increase in chimerism, BM failure, GVHD)
13101	AC→AD	DLI	Y	M	Y (GI, skin, LHS)	50 × 106
14980	AC→AD	DLI	Y	M	N (LHGVHR)	50 × 106
15204	AC→AD	DLI	Y	M	Y (skin, LHS)	50 × 106
16626	AC→AD	CC DLI	Y	M	Y (skin, GI, Liver)	50 × 106
17469	AC→AD	DLI (mega)	Y	M	Y (skin)	150 × 106
18431	AC→AD	DLI	N	L	Y (BM failure)	50 × 106
18860	AC→AD	DLI	Y	M	Y (skin, LHS, GI)	50 × 106
18861	AC→AD	DLI	Y	M	Y (skin, GI, LHs)	50 × 106
19141	AC→AD	DLI	N	L	Y (skin, GI, LHs)	37 × 106

Nine animals that experienced an effect post-DLI. One animal (14980) converted to full donor chimerism without any signs of GVHD, whereas another (17469) converted with skin-restricted GVHD. The other 7 animals developed GVHD. One animal (16626) received a CC (full mismatch to host) DLI after 2 previous AC DLIs and did not have an effect (refer to Table 1). All animals with a GVH response received at least 50 million Tc/eq/kg with the exception of animal 17469, which only converted after receiving 150 million Tc/eq/kg (megadose DLI). No animals received treatment for GVHD. GI indicates gastrointestinal; LHS, lymphohematopoietic system; LHGVHR, lymphohematopoietic GVH response.

Figure 2.

DLI after leukapheresis. Animal 14980 received a DLI on day 218 post-HCT. The animal converted to donor chimerism without GVHD. Animal 19141 received a 37 million CD3⁺ T cell DLI on day 453 post-HCT, and subsequently developed GVHD. Animal 19560 underwent leukaphoresis on day 59 post-HCT and subsequently received a 50 million CD3 T cell DLI. Acute GVHD developed. Animals 17467 and 17469 each received a 50 million CD3 T cell DLI on days 96 and 62 after haplo-HCT, respectively. Animal 17467 had evidence of increased chimerism before succumbing to a non-GVHD-related condition. No effect was observed in animal 17469, and the DLI dose was escalated to 150 million (megadose) CD3⁺ T cell DLI on day 463 post-HCT. The animal converted to full donor chimerism and developed skin-only GVHD.

Sensitized Donor Lymphocytes Fail to Overcome DLI Resistance across MHC Barriers

In an attempt to overcome this resistance to conversion, we assessed whether “sensitized” DLIs could convert tolerant mixed chimeras to full donor chimerism. Two donor animals were sensitized by skin grafting with tissues from host MHC-matched skin before DLI. Animals 13272 and 14376 are 2 tolerant mixed chimeras that initially received unmodified DLIs (PBMC product normalized for CD3 T cells) and were resistant to conversion (Table 1). These animals then received a “sensitized” DLI from a donor that had rejected a skin graft from the host chimera. Donor antichimera CMLs were tested between 90 and 100 days after skin grafting and before DLI. Sensitized donor 14179 (AC) antichimera (14736) CML responses are shown in Figure 3A and compared with antichimera responses of a naive AC animal (Figure 3A). Donor PBMCs showed increased percent specific lysis of chimera PBMCs compared with naive PBMCs. Ten Õ 10⁶ CD3⁺ Tc/kg were delivered to animal 14376 (Figure 3) and 50 × 10⁶ CD3+ Tc/kg to animal 13272. Interestingly, neither recipient animal converted to full donor chimerism or exhibited any BM effect.

Figure 3.

Sensitized and CC DLIs. (A, left) Percent chimerism of animal 14376 after hapioidentical HCT. A 50 million CD3⁺ T cell/kg DLI was delivered on day 36 and a “sensitized” DLI 1 year post-HCT (arrows). (A, middle) CML from a naive AC (used in comparison with the right panel). (A, right) CML from animal 14179 after being sensitized by host MHC-matched skin (to animal 14376, an AD). Assay was performed 92 days after skin grafting. (B, left) Percent chimerism of animal 16626 after haploidentical HCT. The animal was a low level chimera in all lineages. Animal 16626 received a donor AC DLI of 1 million CD3 T cells on day 152 post-HCT followed by a second DLI of 50 million AC donor-matched CD3⁺ T cells/kg (arrows). No effect of either DLI was observed. The third and last DLI was a CC DLI. This DLI was haplo-matched to donor cells (AC) but completely mismatched to the host MHC (AD). (B, middle) CML from chimeric animal 16626 responding to a third-party control while unresponsive to the PBMCs from the CC DLI donor. (B, right) CML using CC effectors and animal 16626 chimeric cells as targets before the DLI.

Overcoming DLI Resistance

In addition to using sensitized donor lymphocytes, 3 other approaches were attempted to overcome DLI resistance, including increasing the dose of DLI, delivering lymphocytes fully mismatched to the original host MHC (AD) but not to the donor MHC (AC), and removing chimeric host peripheral blood cell populations through extensive recipient leukapheresis immediately before DLI.

We attempted an increased DLI dose of 150 × 10⁶ Tc/kg (megadose DLI) in one animal (no. 17469) after having failed to convert after a 50 × 10⁶ Tc/kg DLI. This dose is 3 times higher than the maximum dose delivered to patients in clinical DLI at Massachusetts General Hospital (T. R. Spitzer, personal communication). The dose chosen of 150 × 10⁶ Tc/kg was based on murine studies that demonstrated successful and reliable conversion of stable mixed murine chimeras after receiving a DLI of 150 to 300 × 10⁶ CD3⁺ Tc/kg (after RIC regimens). Mice convert to full donor chimeras without inducing GVHD [27,28]. In our study, animal 17469 successfully converted to full donor chimerism beginning 14 days after receiving a 150 × 10⁶ Tc/kg DLI (Figure 2). Animal 17469 underwent successful immune reconstitution and developed skin-restricted GVHD and exhibited no signs of GVHD in the liver or the gastrointestinal tract.

Resistance to conversion was overcome in another animal that previously failed to convert after 2 DLIs. Animal 16626 received a haploidentical HCT (SLA^a/c→SLA^a/d, also referred as AC→AD) (Tables 1 and 2 and Figure 3B) and 2 AC DLIs of 1 × 10⁶ CD3+ Tc/kg and 50 × 10⁶ CD3+ Tc/kg on days 153 and 566 post-HCT, respectively. In the absence of conversion, a third CC leukocyte infusion was delivered, which was fully mismatched to the host (AD) MHC-I and -II but haplo-matched to the donor (AC) (Figure 3B). At the time of the third DLI, animal 16626 had multilineage chimerism, albeit at low levels, and was tolerant to CC PBMCs in vitro (Figure 3B). In contrast, CC PBMCs had strong cytolytic responses against the host chimera (Figure 3B). The CC DLI dose delivered was equal to the previous AC DLI of 50 × 10⁶ CD3+ Tc/kg. Within 2 weeks, the CC DLI induced conversion to full donor chimerism and acute GVHD (Table 2 and Figure 3), requiring euthanasia of the animal.

Because increased DLI doses of 150 × 10⁶ CD3+ Tc/kg or fully mismatched DLIs are not currently clinically applicable, we attempted an alternative DLI approach. We leukapheresed 4 chimeras (animal nos. 17467, 17469, 19560, and 19561) 2 months after HCT and subsequently delivered a 50 × 10⁶ CD3⁺ Tc/kg DLI immediately after (Figure 2 and Table 3). We aimed to decrease putative regulatory cell populations (which prevent donor antihost proliferative responses in vitro; preliminary data not shown) that, we hypothesize, may be responsible for the resistance to conversion in this protocol. Table 3 summarizes the outcomes of the 4 animals. The total number of PBMCs leukapheresed ranged from 1.2 to 4.5 × 10⁹ PBMCs. All animals were of similar weight at the time of leukapheresis and DLI. The cells removed per kilogram of body weight ranged from 5.45 to 28 × 10⁷ PBMCs/kg. Animal 17467 (Figure 2) had a mild increase in chimerism after leukapheresis and DLI. Animals 19560 and 19561 developed GVHD and subsequently died. The leukapheresis and DLI in animal 17469 did not have an effect. Overall, of the 4 animals that received a leukapheresis and DLI, 2 animals developed GVHD (19560 and 19561), 1 had evidence of peripheral blood chimerism increases before dying of other causes (17467), and 1 animal had no effect (17469).

DISCUSSION

Although stable mixed chimerism across MHC barriers has been reliably achieved in various murine and swine models, translation to nonhuman primate and clinical protocols has been largely unsuccessful. Nevertheless, mixed chimerism, albeit transient, is clinically relevant in the setting of BM transplantation/HCT for the treatment of hematologic malignancies. Furthermore, establishing the mechanisms of DLI resistance in a large animal model of allo-HCT across MHC barriers will surely facilitate the understanding of DLI resistance in the clinical setting. Our model meets the criteria expected for GVHD first postulated by Billingham, which include the presence of immunocompetent cells in the donor inoculum, the inability of the recipient to mount an immune response to the donor cells, and the presence of a histocompatibility difference between the donor and recipient [29]. The immunologic status of tolerant mixed chimeras in this study before DLI satisfied all the requirements for GVHD [29]. Despite our model meeting these criteria, we observed a strong GVH resistance after DLI. In our studies, attempts to abrogate this resistance were based on the hypothesis of a regulatory cell population (regulatory T cells) [30–32] mediating this resistance and trying to either “overpower” it by increasing the DLI dose, using sensitized DLIs and partial MHC sharing with CC lymphocytes, or “removing” it via leukapheresis. In mice, central deletional tolerance and peripheral regulatory mechanisms have been shown to control allo-immune responses [33–36]. Because donor T cells are not depleted and host T cells are partially and transiently depleted in this large animal model [24,25,37], unlike the complete donor and host T cell depletion that can be achieved in rodent studies, we speculate that immune regulatory T cell mechanisms are involved in controlling both donor antihost and host antidonor responses [30,38]. We are currently exploring the mechanism of immunoregulation in this model.

It is important to emphasize that our swine model differs from clinical protocols in 2 important aspects. First, and as previously mentioned, the initial conditioning regimen used in our study is milder than those currently used in clinical HCT. Our regimen results in a transient decrease in host T cells without inducing protracted lymphopenias. At the time of HCT, host WBC counts range from 3000 to 5000 cells/μL and host CD3⁺ T cells range from 300 to 1000 cells/μL (data not shown). WBC counts recover to normal levels within a month post-HCT and during the period of CyA coverage. Second, the donor inoculum is not T cell depleted before transplant and contains an estimation of 1 to 5 × 10⁸ CD3⁺ T cells/kg.

Several conclusions can be drawn from the data in this study. First, our RIC allo-HCT protocol results in a mixed chimeric state without GVHD that is resistant to most unmodified DLIs at a clinical dose. The dose of T cells given as part of the DLI in this study is more than sufficient to cause GVHD in myeloablatively conditioned recipients. Haploidentical HCT at T cell doses 50 to 100 times lower than the DLI dose delivered in this study, when delivered in myeloablatively conditioned swine, consistently results in severe, acute GVHD [39]. Thus, of 32 DLIs given at a clinical dose of 1 to 50 million T cells (and 1 megadose DLI), the lack of any effect in 21 cases was striking. In previous work using a canine model of mixed chimerism, sensitized lymphocytes were able to overcome this resistance and induce conversion across minor histocompatibility barriers in 8/8 canines [40]. When attempted in our swine haploidentical model across MHC barriers, sensitized lymphocytes did not recapitulate what was observed in the minor histocompatibility canine studies, likely because of a potent host-versus-graft protection.

Second, our data presents proof of concept that a very high T cell doses and/or leukodepletion are necessary to potentially overcome the resistance to conversion. In our study, the weight of animals receiving a DLI ranged from 25 to 60 kg. A 50-kg swine with a WBC count of 15,000 would have in the order of 18 to 20 × 10⁹ circulating CD3⁺ T cells. A 50-kg animal receiving a 50 × 10⁶ Tc/kg DLI would receive a total dose of 2.5 × 10⁹ CD3+ Tc. This represents one-eighth to one-tenth of the total circulating T cells. Based on our results, it is likely that tolerant mixed chimeras may require higher doses of naive T cells to induce a consistent GVH response. Only animals that received the higher doses (≥50 million Tc/kg), with the exception of 1 animal (no. 19141 received a DLI dose of 37 × 10⁶ Tc/kg), exhibited a response to DLI. Leukapheresing animals before DLI resulted in greater rates of conversion and GVHD. We hypothesize that removal of putative regulatory cells (potentially T regulatory cells) and the inflammation induced by the leukapheresis created an environment favoring the infused naive donor T cells. Mouse models reliably achieved conversion after DLIs of 150 to 300 × 10⁶ Tc/kg [27,28,41]. Higher doses (>300 × 10⁶ Tc/kg) in mice cause GVHD (M. Sykes, personal communication). The 1 animal (no. 17469) that resisted leukapheresis and DLI actually converted to full donor chimerism after a 150 × 10⁶ Tc/kg DLI. We do acknowledge the limited numbers presented, and more animals are planned to be studied to elucidate the responsible cell populations for the resistance of conversion in this model.

Third, conversion to full donor chimerism is often associated with the development of GVHD. To minimize GVHD, all chimeras received the DLIs long after the preparatory regimen to ensure minimal residual inflammation secondary to the transplant or conditioning regimen. Billiau et al. [42] demonstrated that the most potent GVH responses in murine studies occurred when DLI was administered 3 weeks post-HCT and when donor chimerism was between 60% and 70%. In Billiau et al.’s study, if DLIs were delivered at week 12 post-HCT and at a time when donor chimerism was 90%, neither GVHD nor graft-versus-leukemia effects were observed. However, when chimerism levels were <50% and DLI was given, GVHD occurred. These studies imply that a minimal amount of host antigen presenting cells may be required via direct presentation to provide sufficient antigen to donor T cells and induce the GVH responses [27,28,41,42]. However, if too many host antigen presenting cells are present at the time of DLI, GVHD can develop. In our study, most animals that received a DLI were low chimeras (<50%), and based on murine studies, GVHD would likely be observed. This hypothesis was accurate because most swine chimeras that converted after DLI and developed GVHD had low (<50%) chimerism. The 1 animal that converted without GVHD had chimerism levels of 70%.

Fourth, partial MHC sharing may be important for controlling GVH responses. Supportive evidence for the control of GVH responses through shared MHC was demonstrated in animal 16626. After failing to convert to full donor chimerism after 2 haplo-matched AC DLIs, the chimera converted upon delivery of a full mismatched CC DLI at an equivalent dose. Hence, based on several publications and previous swine studies by Sonntag et al. and Leguern et al. [30,32], we could speculate that GVH and host-versus-graft responses may have been controlled by regulatory T cells through the peptide major histocompatibility complex-T cell receptor interactions (pMHC-TCR) interactions directed to SLA-a that was shared by the AC (donor) and AD (recipient). Absence of SLA-a in our CC DLI may have prevented/decreased optimal regulatory T cell function/protection. Further studies are needed to identify the cell populations responsible for these protective effects.

The clinically most promising result of our study was that in isolated cases, DLI was able to convert a tolerant mixed chimera to full donor with minimal or no GVHD. Although these effects have been reported in mice, overall, our experience in miniature swine more similarly mirrors the outcomes reported in humans, in which conversion to full donor chimerism is associated with severe GVHD [27]. Although the exact mechanism responsible for overcoming this resistance to conversion in our protocol is still unclear, there is evidence for a strong immune regulatory mechanism induced and maintained through the RIC-HCT protocol resulting in long-lasting memory cells capable of controlling GVH responses. Resistance to conversion was overcome when these cells were removed from the host chimera through leukapheresis or “overpowered” by increasing the DLI dose and increasing the number of alloreactive donor T cells.

In summary, resistance to conversion after DLI and protection from GVHD is strong in tolerant mixed chimeras. Overcoming this resistance either through extensive leukapheresis of the host chimera immediately before DLI or using very high T cell doses is usually associated with GVHD. However, in some isolated cases, conversion without GVHD can be induced with our protocol. The use of DLI in our model and the mechanisms responsible for the resistance to conversion need to be further studied to facilitate clinical application. We believe that understanding these processes in the swine model will facilitate the understanding of DLI resistance in the clinical setting.