Evaluation of the tolerogenic effects of donor bone marrow cells using a severe combined immunodeficient mouse-human islet transplant model

Abstract

The immunoregulatory role of human donor bone marrow cells (DBMC) has been studied extensively in our laboratory using in vitro and ex vivo assays. However, new experimental systems that can overcome the limitations of tissue culture assays but with more clinical relevance than purely animal experimentation, needed to be generated. Therefore we have developed a new human peripheral blood lymphocyte (PBL) severe combined immunodeficient (SCID) mouse islet transplantation model without the occurrence of graft-versus-host disease (GvHD) and have used it to evaluate the tolerogenic effects of DBMC. Nonobese diabetogenic (NOD)-SCID mice were transplanted with human deceased donor islets and were reconstituted with human PBL (allogeneic to islets; denoted as recipient) with or without DBMC from the islet donor. It was observed that the most cellularly economical dose was 3000 islets per animal and that injection into the portal vein was better than implantation under the kidney capsule. Even though maximal lymphoid reconstitution was observed with 40-million fresh and anti-CD3 activated recipient PBL (conventional method), the mice developed severe graft GvHD. However, with the new method of reconstitution where animals were injected with 20-million anti-CD3-activated plus 40-million anti-donor-activated recipient PBL, no discernible GvHD was observed. More importantly, this latter method was associated with islet transplant rejection, which in turn could be abrogated by co-injection of the animals with DBMC. These in vivo results confirmed our previous in vitro observations that human DBMC have regulatory activity.

Figure 1

Reconstitution of NOD-SCID Mice with human PBL (n = 4/group). NOD-SCID mice were injected according to the following schedule in various groups: (Group I) on day 0 with 20 × 10⁶ fresh recipient PBL, and on day 3 with 20 × 10⁶ anti-CD3 activated recipient PBL, (Group II) on day 0 with 20 × 10⁶ fresh, and on day 3 with 20 × 10⁶ anti– donor-activated recipient PBL, and (Group III) on day 3 with 20 × 10⁶ anti-CD3 activated recipient PBL and on days 5 and 7 each with 20 × 10⁶ anti– donor-activated recipient PBL. On indicated days, tail vein blood was obtained from the mice and flow cytometric analysis was performed for the indicated phenotypes gated on total leukocytes based on forward versus side scatter. The standard deviations (SD; not shown) were less than 10%. A representative experiment for one of the mice on day 21 in group III is shown in histograms IIIA-IIIF so as to further depict the characteristics of the human cells that were gated (IIIA) on those that did not bind by mouse CD45-FITC.

Figure 2

Human Immunoglobulin production in NOD-SCID mice reconstituted with human PBL: NOD-SCID mice were injected according to the following schedule in various groups: (Group I) on day 0 with 20 × 10⁶ fresh and on day 3 with 20 × 10⁶ anti-CD3 activated recipient PBL, (Group II) on day 0 with 20 × 10⁶ fresh and on day 3 with 20 × 10⁶ anti– donor-activated recipient PBL, and (Group III) on day 3 with 20 × 10⁶ anti-CD3 activated recipient PBL and on days 5 and 7 each with 20 × 10⁶ anti– donor-activated recipient PBL, as in Figure 1. The plasma collected from the venous blood from the three groups of mice on the indicated days was used to estimate the level of human immunoglobulins (Ig) by a sandwich ELISA. The data are shown as mean ± SD with four animals in each group.

Figure 3

Survival of human islets in NOD-SCID mice. Indicated numbers of purified human islet equivalents were transplanted into the portal vein or under the kidney capsule of NOD-SCID mice. C-peptide levels were monitored in serial plasma samples at weekly intervals as indicated using a human C-peptide RIA kit. The data are shown as mean ± SD with five animals in each group.

Figure 4

Rejection of human islets transplanted into NOD-SCID mice in the presence of anti-donor human PBL. NOD-SCID mice were (Group IV) transplanted with 3000 human islets on day 0, (Group V) transplanted with 3000 human islets on day 0 and then reconstituted on day 3 with 20 ×10⁶ anti-CD3 activated recipient PBL, then on days 5 and 7 each with 20 ×10⁶ recipient PBL activated with donor irradiated spleen cells, and (Group III) reconstituted on day 3 with 20 ×10⁶ anti-CD3 activated recipient PBL, then on days 5 and 7 each with 20 ×10⁶ recipient PBL activated with donor irradiated spleen cells (without islet transplantation). The plasma samples were collected on indicated days and the levels of secreted human C-peptide were estimated by a competitive RIA. The data are shown as mean ± SD with four animals in each group.

Figure 5

Inhibition of human islet transplant rejection by NT-LP/DBMC in NOD-SCID mice. NOD-SCID mice were divided into three groups: (1) Group IV, negative control group, transplanted intraportally with 3000 human islets on day 0 and then injected with PBS only and therefore had no reconstitution; (2) Group V, positive control group, transplanted intraportally with 3000 human islets on day 0 and then reconstituted by injection on day 3 with 20 ×10⁶ recipient PBL activated with anti-CD3 and co-cultured with 10 ×10⁶ donor irradiated spleen cells for 3 days, followed, on days 5 and 7, with 20 ×10⁶ recipient PBL co-cultured with 30 ×10⁶ donor irradiated spleen cells for 5 or 7 days; and (3) Group VI, experimental group, transplanted intraportally with 3000 human islets on day 0 and then reconstituted by injecting on day 3 with 20 ×10⁶ recipient PBL activated with anti-CD3 and co-cultured with 10 ×10⁶ NT-LP/DBMC for 3 days, followed by on days 5 and 7 each with 20 ×10⁶ recipient PBL co-cultured with 20 ×10⁶ donor irradiated spleen cells plus 10 ×10⁶ NT-LP/DBMC for 5 or 7 days. Then, plasma samples were collected on indicated days and the levels of secreted human C-peptide levels were measured by a competitive RIA. The data are shown as mean ±SD with five animals in each group.