Alloreactivity following in utero transplantation of cytokine-stimulated hematopoietic stem cells

Experimental Hematology 30 (2002) 617–624

Alloreactivity following in utero transplantation of cytokine-stimulated hematopoietic stem cells: The role of recipient CD4 cells Hassan Sefriouia, Jody Donahuea, Anand Shanker Srivastavaa, Elizabeth Gilpinb, Tzong-Hae Leec, and Ewa Carriera a Department of Medicine and Pediatrics, bFamily and Preventive Medicine, University of California, San Diego, San Diego, Calif., USA; Department of Laboratory Medicine and Irwin Memorial Blood Center, University of California, San Francisco, San Francisco, Calif., USA

c

(Received 29 November 2001; revised 5 February 2002; accepted 21 February 2002)

Objective. We have previously reported immunity to donor antigens following in utero transplantation (IUT) of cytokine-stimulated allogeneic hematopoietic stem cells (sca/lin) (day 9 of gestation). Transplanted mice showed accelerated rejection of donor skin grafts and high anti-donor cytotoxic response, a finding not seen in the control mice. This was accompanied by an enhancement of Th1 over Th2 cytokine production and persistent donor microchimerism [1,2]. In order to assess the role of the thymus in allograft rejection, prenatal transplants were performed under similar experimental conditions at a later gestational age, when the thymus is more developed (day 13). Materials and Methods. Cytokine-stimulated stem cell factor (SCF) and granulocyte colonystimulating factor (G-CSF)–purified sca/lin cells of C57BL/6 (H-2b, 1E) background were injected into MHC-mismatched BALB/c (H-2d, 1E) fetal mouse recipients at day 13 of gestation. Chimerism was determined by highly sensitive (0.001%) semiquantitative polymerase chain reaction (PCR) [3,4]. Mixed lymphocyte reaction (MLR) and cytotoxic T-cell assay (CTL) were used to evaluate tolerance vs immunity. Cytokine levels were quantified in MLR supernatants using ELISA assay. The percent of T cells was determined by flow cytometry (FACS) and CD4/CD8 ratio calculated. Postnatal boosts (transplants without conditioning) were performed at 6 months of age to enhance donor chimerism and test the degree of tolerance. Results. When assayed at 4 months of age, donor-type cells were not detected in the spleen or in the peripheral blood of BALB/c mice inoculated with C57BL/6 sca/lin cells. Transplanted but not control animals demonstrated high anti-donor MLR but not CTL responses. The increase of MLR reactivity was correlated with high levels of IL-2. Furthermore, transplanted mice showed higher resistance to postnatal boosts with allogeneic bone marrow (BM) cells, when compared to the control mice. The later resistance was accompanied by the expansion of host-type CD4 cells. Conclusion. These data demonstrate that transplantation of cytokine-stimulated sca/lin allogeneic cells at 13 days of fetal development leads to the allosensitization, characterized by an enhancement of MLR alloreactivity and by the rejection of postnatal boosts (transplants without conditioning). Host-type CD4 cells might play a central role in this rejection. These findings indicate that the late injection of allogeneic cells may result in the development of allosensitization with subsequent donor graft rejection. Precise conditions for the development of tolerance must be established before prenatal transplants in humans with conditions other than SCID can be done. © 2002 International Society for Experimental Hematology. Published by Elsevier Science Inc.

Offprint requests to: Ewa Carrier, M.D., University of California San Diego Blood and Marrow Transplantation Division, 9500 Gilman Drive, 0062, La Jolla, CA 92093-0062; E-mail: [email protected] This work was presented at the 30th Annual Meeting of the International Society for Experimental Hematology, Aug 25–28, 2001, Tokyo, Japan.

0301-472X/02 $–see front matter. Copyright © 2002 International Society for Experimental Hematology. Published by Elsevier Science Inc. PII S0301-472X(02)0 0 8 0 3 - 2

618

H. Sefrioui et al./Experimental Hematology 30 (2002) 617–624

In utero transplantation (IUT) has been effective in creating postnatal tolerance to solid allograft transplantation [5,6] and in treating patients with defective immune systems [7–10]. However, the use of IUT for patients with intact immune systems, i.e., inborn errors of metabolism and hemoglobinopathies, has been discouraging [11,12]. The low degree of tolerance and donor chimerism in these cases remains a major barrier preventing successful clinical outcome. These barriers are 1) robust fetal hematopoiesis and 2) immunological responses of the actively developing fetal immune system. Our hypothesis is that the second-trimester fetal immune system is capable of generating an immune response that can lead to the rejection of postnatal transplants. Recently, the classical paradigm of the prenatal and neonatal tolerance window was challenged by several studies showing that fetuses may evoke immune responses [13]. For example, low levels of mature T cells were detected at an early stage of human fetal development [14–17]. Development of immunity rather than tolerance following IUT of human or murine allogeneic hematopoietic stem cells following IUT has been previously shown [18–24]. Other studies demonstrated that murine fetal natural killer (NK) cells can be highly proliferative and secrete cytokines similar to what is seen in an adult system [25,26]. Furthermore, it was shown that trophoblast expression of Fas-L had potential lymphocyte targets, i.e., Fas-expressing maternal immune cell [27]. We have previously demonstrated induction of tolerance following prenatal transplantation with nonstimulated allogeneic fetal liver cells [3]. Recently, we have reported development of prenatal sensitization (day 9 of gestation, prethymic, first trimester) following IUT of cytokine-stimulated allogeneic sca/lin hematopoietic stem cells [2]. Transplanted but not control mice had accelerated donor skin graft rejection, increase in anti-donor cytotoxicity, and enhancement of Th1 cytokine production. We hypothesized that cytokinestimulated sca/lin cells contained antigen-presenting cells that contributed to the development of allosensitization. Additionally, we have also hypothesized that persistent low levels of donor chimerism maintained this immunity. To determine whether inoculation of allogeneic sca/  lin cells at later gestational age had similar influences on the outcome of microchimerism and tolerance vs immunity, we injected sca/lin cells at day 13 of gestation (postthymic, second trimester), when the immune system is more developed. Immunological studies and postnatal boosts (transplants without conditioning) with donor cells were performed to determine induction of tolerance/immunity.

Materials and methods Mice Six- to eight-week-old C57BL/6 (GPI-1b, Thy1-2, H2-b 1-E), BALB/c (GPI-1a, Thy1-2, H2-d 1-E), and third-party C3H (GPI-

1b, Thy1-2, H-2k) mice were purchased from Simonsen Laboratory (Gilroy, CA, USA) and bred in our colony. Mice were housed in a group of 3 to 4 mice/cage under filter caps in the Animal Care Facility at the University of California, San Diego (UCSD). The Animal Care Facility at UCSD is a federally approved, pathogenfree facility with 24-hour-per-day veterinary care. All experimental procedures were reviewed by the Committee on Animal Research at UCSD. Sca/lin cell purification Six- to eight-week-old male donor C57BL/6 mice were injected subcutaneously for 6 days with recombinant human granulocyte colony-stimulating factor (rhG-CSF: 200 g/kg/day) and pegylated recombinant rat stem cell factor (PEG-rrSCF: 25 /kg/day) to mobilize hematopoietic stem cells (HSC) [28]. Seven days after injection, mice were sacrificed, and splenocytes were collected and passed through a nylon-mesh 40-m cell strainer. After washing, viable cells were separated by density gradient centrifugation (Histopaque-1083, Sigma-Aldrich, St. Louis, MO, USA) and Fc receptor blockade was performed using 5% rabbit serum in RPMI. Sca/lin cells were selected using sca-1 multisort kits (Milteny Biotec, Auburn, CA, USA). Briefly, cells were incubated with sca beads for 25 minutes at 4C, washed three times in phosphate-buffered saline (PBS) with 0.5% bovine serum albumin (BSA), and passed through a magnetic field. Positively selected cells were detached from the microbeads and incubated for 10 minutes at 4C with a biotin-labeled antibody cocktail of lineage-specific markers (CD11b, Lys-G, Ter119, CD4, CD8, CD45/B220, Pharmingen, San Diego, CA, USA) and stop solution. After washing, cells were incubated at 4C with streptavidin microbeads. Ten minutes later, they were passed through a magnetic column to collect the negative fraction and examined for purity using flow cytometric analysis. Purified sca/lin cells (95%) were tested for their multilineage proliferative capacities in methylcellulose cultures. Similar to what we have previously reported [2], sca/lin cells showed an enrichment of CFU-mix, BFU-E, and CFU-GM colonies. In utero transplantation The in utero procedure was performed as previously described [3]. Briefly, micropipettes were prepared from capillaries with a Brown and Fleming micropipette puller (T-87, Sutter Instruments, San Rafael, CA, USA). The end of the needle was cut under a microscope with jeweler’s forceps, and the tip was sharpened on a micro-grinding wheel. The day on which the vaginal plug was identified was considered as day 0 of gestation. Thirteen days later, pregnant mice were anesthetized and the uterus exposed through a ventral laparotomy. The tip of the needle was inserted through the uterine wall into the peritoneal cavity of each fetus and 9.5  109 sca/lin cells per kg fetal weight (5  104 cells/fetus) were injected. Polymerase chain reaction assay DNA samples were prepared and processed as previously described [3,4]. Briefly, 100 L of peripheral blood was subjected to saponin lysis buffer (0.4% saponin [ICN Biochemical, Cleveland, OH, USA] in 0.5% NaCl). Fifty L of polymerase chain reaction (PCR) solution B (10 mmol/L Tris, pH 8.3; 2.5 mmol/L MgCL; 1% Tween-20; NP40; 0.4 mg/mL proteinase K) was added to each leukocyte pellet. Preparations were incubated at 60C for 1.5 hours with vortexing every 20 minutes followed by 95C for 2 hours. After amplification, the preparations were hybridized at 95C for 5

H. Sefrioui et al./Experimental Hematology 30 (2002) 617–624

minutes and 60C for 5 minutes with [32P]-labeled oligonucleotide probes. The DNA lysate was tested in duplicate in a single PCR polyacrylamide gel electrophoresis/autoradiography and run together with duplicate standard dilution series. Standard curves composed of DNA derived from 10-fold serial dilutions of donor (male) in recipient (female) cells (103, 102, 101, 100) were analyzed in parallel with the samples (PCR primers used for Y chromosome are: GAG AGG CAC AAG TTG GCC C and GGC TTT TCC ACC TGC ATC C). The gel was exposed to the XAR-5 autoradiographs film (Eastman Kodak Co., Rochester, NY, USA) with an enhancing screen for 2 hours and left overnight at room temperature. Autoradiographs were analyzed with the Millipore Bio-Image electrophoresis analyzer with the whole-blood analyzer application software (Millipore, Ann Arbor, MI, USA). The reproducibility of the image analysis system was assessed by replicate (5) scanning of 23 sample data points spanning the assay’s dynamic range 12 to 500 cells: the mean coefficient of variance was 8.320/ 0, with a range of 3.70/0 to 17.80/0. Based on analysis of in vitro dilutions of male cells into the female blood, the assay detects as few as 1 donor cell per 25 L (1.5–2  105 cells) each of recipient blood (the input volume per PCR). Hemisplenectomy Mice were anesthetized with inhalation anesthesia isoflurane. The surgical site was rinsed with alcohol, and a small incision made in the dorsal abdominal cavity just below the spleen. One-quarter of the spleen was separated by isolating the tissue section with suture to cauterize the blood vessels, then cutting the tissue section from the remainder of the spleen mass. The incision was closed with 5-0 chromic gut sutures, and mice were monitored until fully recovered. Excised tissue was placed in 35-mm petri dishes with RPMI and brought into a single-cell suspension by first teasing the tissue mass between two sterile, frosted microscope slides, then filtering the resultant suspension over 50-m nylon mesh via a 500-L pipette tip. Mixed lymphocyte reaction (MLR) Responder spleen cells were collected and suspended in culture media (RPMI-1640, 100 U penicillin/streptomycin, 10% fetal bovine serum [FBS], 25 mM HEPES buffer, 0.09 mM NEAA, 1 mM sodium pyruvate, and 50 M -mercaptoethanol). Briefly, 2  105 responder cells were cocultured with 2  105 irradiated stimulator spleen cells at 37C with a 5% CO2-to-air mixture. Following 5 days in culture, 1 Ci H3 thymidine per well was added. Cells were harvested after 24 hours of additional incubation. Tritiated thymidine incorporation was expressed as counts per minute from which the counts of thymidine alone were subtracted. Cytotoxic T-cell assay (CTL) Cytotoxicity was prepared according to our standard procedures [2]. Responder spleen cells (6  106/mL) were cocultured with 2500 rad irradiated spleen cells (stimulator cells) (6  106/mL) in 24-well tissue culture plates at 37C with a 5% CO2-to-air mixture. Culture media was composed of RPMI-1640, 100 U penicillin/ streptomycin, 10% FBS, 25 mM HEPES buffer, 0.09 mM NEAA, 1 mM sodium pyruvate, and 50 M -mercaptoethanol. After 5 days of incubation, cells were harvested and aliquoted into triplicate 96-well tissue culture plates. Responder cells were cultured with 51Cr (NEN Life Science Products, Boston, MA, USA)-labeled tumor cell lines: J558L (BALB/c), EL-4 (C57Bl/6), and R1.1 (C3H). Four hours later, 100 L of supernatant were transferred

619

into a 96-well sample plate (Wallac, Turku, Finland), 100 L of scintillation fluid was added, and the plate was sealed. A Wallac Oy 1450 Microbeta counter was used to determine  emission. Results are expressed as the percent of specific lysis and were calculated by standard methods: (experimental release  spontaneous release)/ (maximum target release  spontaneous target release)  100. NK cell activity NK activity was measured by culturing responder cells together with the YAC-1 cell line (ATCC, Manassas, VA, USA) as target T cells using the procedures previously established in our laboratory [1]. Spleen cells were washed and challenged with YAC-1 cell line at varying effector-to-target ratios in a 96-well round-bottomed tissue culture. Plates were incubated in humidified 37C and 5% CO2 for 4 hours. Cytotoxicity was measured by LDH release with the Cytotox 96 (Promega) nonradioactive cytotoxicity kit. Cytokines quantification Interleukin (IL)-2, IL-4, and interferon (IFN- ) were quantified using standard ELISA assay performed by M. Croft from the La Jolla Institute of Allergy and Immunology. Culture plates were coated with primary unlabeled antibody in carbonate buffer and incubated overnight at 4C. Plates were blocked with 1% BSA, 5% fetal calf serum (FCS), and 0.05% Tween-20 in PBS solution. Standards (Pharmingen, San Diego, CA, USA) and MLR supernatants were added to the blocking solution and plates were incubated for 2 hours at room temperature. Biotin-labeled antibody in blocking solution was added to each well. To enhance detection and specificity, extravidin, anti-avidin biotin, and extravidin–horseradish peroxidase antibodies were added and separated by vigorous washing. Urea hydrogen peroxidase and o-phenolenediamine dihydrochloride in citrate phosphate buffer was added as a substrate. Plates were incubated for 20 minutes at room temperature and the reaction was stopped with 25% H2SO4. Cell preparation and flow cytometry Peripheral blood was drawn from the lateral tail vein and placed in a falcon tube. Ammonium chloride lysis solution was added to lyse red blood cells. Samples were washed twice with PBS and incubated with CD16/CD32 Fc block for nonspecific antibody binding and kept at 4C for 45–60 minutes. After washing, propidium iodide was added to exclude dead cells from analysis. Immunophenotyping was performed by staining with phycoerythrin (PE)-conjugated anti-CD45, CD3, H-2d, and hamster-IgG, or fluorescein isothiocyanate (FITC)-conjugated anti-H-2b, CD4, CD8, and ratIgM, or APC-conjugated anti-CD45 monoclonal antibodies (Pharmingen). Three-color flow cytometry analysis was performed on a FACScan (Becton-Dickinson, San Jose, CA, USA). Postnatal boosts (transplants without conditioning) BALB/c mice transplanted in utero with C57BL/6 sca/lin hematopoietic stem cells and age-matched nontransplanted controls received boosts at 6 months of age. Injections consisted of 40  106 adult C57BL/6 bone marrow (BM) cells infused at 4 consecutive days via the tail vein. BM cells were obtained from 6- to 8week-old mice by flushing the femora and tibia with RPMI containing heparin. Cells were strained through a nylon-mesh 40-m cell strainer, washed in RPMI, and counted in acridine orange. Single-cell suspension was prepared and cells were injected in 200 L PBS within 6 hours of preparation.

620

H. Sefrioui et al./Experimental Hematology 30 (2002) 617–624

Statistical analysis An analysis of variance (ANOVA) and t-test were used for statistical analyses. A p-value of less than 0.05 was considered statistically significant. Results are expressed as mean standard deviation. ANOVA was used to compare results in more than two groups of mice. When tests were performed repeatedly at different dilutions, etc., repeated ANOVA measures were utilized and main effects and interactions were tested. When only two groups of mice were compared, a t-test was used.

Results Cytotoxic T-cell and NK assays We have previously demonstrated that transplantation of cytokine-stimulated murine allogeneic sca/lin cells at day 9 of fetal development leads to immunity rather than tolerance [2]. This immunity was correlated with an enhancement of anti-donor cytotoxicity, accelerated rejection of donor skin grafts, and suboptimal response to the postnatal boosts as compared to controls. It was our hypothesis that the persistent low level of donor chimerism maintained this state of immunity. In order to determine whether injections performed at a later gestational age, when thymus is more developed, affects donor chimerism and tolerance, we transplanted cytokine-stimulated C57BL/6 sca/lin cells into the 13-day-old BALB/c fetuses (C57BL/6 →BALB/c). Thereafter, anti-donor CTL response was performed at 4 months of age. Briefly, after in vivo hemisplenectomy, spleen cells from C57BL/6 →BALB/c (n  11) and control BALB/c (n  6) mice were cultured together with irradiated spleen cells for 5 days in vitro, and tested with 51Cr-labeled target cells. As shown in Figure 1, transplanted animals displayed a significant decrease in lytic activity against donor-type C57BL/6 target cells, when compared to the control BALB/c mice (repeated ANOVA main effect, p 0.01). In contrast, CTL activity against third-party (C3H) and syngeneic (BALB/c) target cells were similar (p 0.376) in the transplanted and control mice. NK activity tested at the same time point as CTL assay and at 20% lysis was not statistically different between the two groups of mice (ANOVA, p 0.399) (Fig. 2).

Mixed lymphocyte reaction We next examined whether the decrease of anti-donor CTL activity in transplanted mice (Fig. 1) correlates with the decrease in anti-donor MLR responses. For that purpose, spleen cells from 4-month-old C57BL/6 →BALB/c and control BALB/c mice were cultured with the irradiated host, donor, or third-party (C3H) spleen cells. Four days later, 3 H-labeled thymidine was added to measure proliferation. Figure 3 shows one out of two experiments with similar results, demonstrating that anti-donor MLR reactivity was significantly higher in transplanted as compared to the con-

Figure 1. Responsiveness of 4-month-old control BALB/c (n  6) and C57BL/6 ›BALB/c (n  11) mice to the host, donor, and third-party (C3H) antigens as assayed by cytotoxic assay (CTL). Effector:target experiments were performed in triplicate. Anti-donor response of transplanted and control mice was significant, p 0.01. The anti–third party responses between transplanted and control mice were not significant, p 0.376.

trol animals (t-test, p 0.017). MLR responses against host and third-party antigens were similar and statistically not different.

Cytokine release In order to correlate T-cell proliferative responses to alloantigen with cytokine release, MLR supernatants were collected and IL-2, IL-4, and IFN- were quantified using ELISA assay. As shown in Figure 4, IL-2 concentration was significantly higher in anti-donor MLR supernatants in transplanted but not control mice (t-test, p 0.012). IFN- and IL-4 levels were similar in the two groups of mice. In anti-host MLR supernatants, IL-2, IFN- , and IL-4 concentrations were low (Fig. 4). This finding is compatible with the cytokine profile shift toward Th1, suggesting allosensitization.

H. Sefrioui et al./Experimental Hematology 30 (2002) 617–624

621

Figure 2. NK activity against YAC cells of 4-month-old control BALB/c (n  6) and C57BL/6 →BALB/c (n  11) mice. Effector:target experiments were performed in triplicate. No difference was detected between transplanted and control mice.

Chimerism and postnatal boost Donor chimerism in the peripheral blood was tested at monthly intervals starting at 2 weeks of age and in the spleen and the liver at 6 months of age and postmortem. When assayed at 6 months of age or before that (data not shown), C57BL/6 →BALB/c mice had no evidence of donor cells either in the spleen or in the peripheral blood, as demonstrated by highly sensitive, semiquantitative PCR (0.001%) (Fig. 5). Transplants without any conditioning and with high doses of donor bone marrow cells (40  106 cells/day for 4 days) were done at 6 months of age in order to enhance the degree of donor chimerism. When assayed at 2 weeks

Figure 3. Responsiveness of 4-month-old control BALB/c (n  3) and C57BL/6 ›BALB/c (n  3) mice to host, donor, and third-party (C3H) antigens as assayed by mixed lymphocytes culture (MLR). Experiments were performed in triplicate. *p 0.017 when compared to BALB/c mice (1 out of 2 similar experiments).

Figure 4. IL-2, IFN- , and IL-4 levels in 5-day anti-host or anti-donor MLR culture supernatants of control and transplanted BALB/c mice (n  5–6), *p 0.012 when compared to control BALB/c mice.

postboosts, microchimerism was significantly lower in the transplanted than in the control mice (t-test, p 0.005): three of six boosted C57BL/6 →BALB/c mice were not engrafted (Fig. 5). In contrast, four out of four control BALB/c mice never exposed to the donor cells demonstrated significantly higher levels of microchimerism (p 0.005. These data indicate that prenatal exposure to the allogeneic cells may induce memory responses, preventing successful postnatal transplants. CD4/CD8 ratio Transplanted but not control animals demonstrated higher resistance to the postnatal transplants (transient donor chimerism following donor cell injections), as they showed low levels of donor chimerism after such procedure. Additionally, transplanted mice demonstrated a higher anti-donor MLR activity, which could be related to the CD4 T cells. In order to prove this, we determined the percentage of CD4 and CD8 cells before and after the boosts. Figure 6 shows the percentage of CD4 and CD8 cells at 2 weeks postboosts. Boosted C57BL/6 →BALB/c mice, exposed to the donor cells prenatally, demonstrated significant increases in the CD4 levels when compared to the agematched control BALB/c (not transplanted in utero) or nonboosted (C57BL/6 →BALB/c) animals transplanted in utero

622

H. Sefrioui et al./Experimental Hematology 30 (2002) 617–624

Figure 5. Percentage of donor cells in the peripheral blood in C57BL/c →BALB/c mice (n  6) and control BALB/c (n  4) animals before and 2 weeks following postnatal boosts with donor bone marrow cells. Boosts were done at 6 months of age. Engraftment was established by semiquantitative PCR (sensitivity 0.001%). *p 0.0053 when compared to the control boosted BALB/c mice.

(ANOVA, p 0.01) (Fig. 6). This response was associated with the increased clearance of donor cells from the peripheral blood in transplanted mice (Fig. 5). The CD8 levels were similar in the three groups of mice; CD4/CD8 ratio was higher in the transplanted mice when compared to control animals (ANOVA, p 0.01) (Fig. 6). These differences between CD4 and CD8 cells explain different responses of host spleen cells in CTL and MLR cultures and indicate the role of CD4 cells in the subsequent graft rejection. Graft-vs-host disease All mice were observed clinically throughout their lives (24 months) and were weighed at monthly intervals. No evidence of graft-vs-host disease (GVHD)—weight loss, diarrhea, hair loss)—was detected in transplanted or control mice either before or after the postnatal boosts. The in utero transplanted mice remained active and vigorous, ate well, and had normal-appearing skin when compared to the control animals. Discussion Hematopoietic stem cell transplantation (HSCT) has been used as a treatment option for a wide variety of disorders including malignancies, inborn errors of metabolism, and congenital hemoglobinopathies [28]. Despite the fact that the field of HSCT has expanded during the last decade, several problems persist, including graft rejection, GVHD, and toxicity resulting from myeloablative regimens. Recently, in utero transplantation has emerged as an alternative treatment modality for diseases diagnosed prenatally [29,30]. The hypothesis and dogma of the field was that preimmune fetuses would develop tolerance to the allogeneic cells and allow correction of the underlying disorder. To date, approximately 40 prenatal transplants have

Figure 6. Percentage of CD4 and CD8 cells in the peripheral blood CD3 population in boosted (n  5) and nonboosted in utero transplanted BALB/c mice (n  2) and in boosted control, not transplanted BALB/c animals (n  3). Mice were assayed at 2 weeks postboosts. *p 0.01 when compared to control BALB/c or nonboosted C57BL/c ›BALB/c mice.

been performed worldwide, with clinical success demonstrated only in the fetuses with severe combined immunodeficiency syndrome (SCID) [7–9]. In fetuses with intact immune systems, such as congenital hemoglobinopathies and inborn errors of metabolism, no clinical success has been demonstrated so far [11,12]. Chimerism was low and transient, and no improvement in clinical symptoms was found. In some cases, evidence of allosensitization was documented [18,19]. Although IUT has been successfully used for the induction of tolerance to solid organ transplants [5,6], occasionally IUT leads to the development of immunity rather than tolerance to donor antigens [18–24]. This immunity may play a central role in the prenatal allograft rejection and could lead to the rejection of postnatal transplants or even the development of late autoimmunity. We have previously demonstrated allosensitization after early IUT of cytokine-stimulated sca/lin hematopoietic stem cells at (day 9 of gestation) [1,2]. Transplanted but not control mice had accelerated skin allograft rejection, en-

H. Sefrioui et al./Experimental Hematology 30 (2002) 617–624

hanced CTL alloresponses, and a shift toward Th1 cytokine profile. The injected cell population was 90% pure and therefore may have contained progenitors capable of antigen presentation and subsequent rejection. Additionally, at this early gestational age thymic rudiments are already present and could allow development of memory response [31]. It is our hypothesis that an underdeveloped immune system allowed development of a very low-level donor chimerism which maintained allosenitization state, but was insufficient for the development of tolerance. In order to investigate whether delayed (“postthymic,” second trimester) IUT of cytokine-stimulated sca/lin cells has a similar immune effect, these cells were inoculated at day 13 of gestation, when the colonization of thymus by T cells has occurred and TCR receptors on fetal thymocytes are developed [32]. In our present model of IUT, enhanced MLR but decreased CTL responses were observed with increased CD4/ CD8 ratios and IL-2 production. The finding that anti-donor MLR, but not CTL responses, increased following transplantation was previously reported [33–35] and was related to the expansion of CD4 cells. Similarly, we have observed significant expansion of CD4 cells but not CD8 cells upon reexposure to the donor cells during the postnatal boosts (Fig. 6). These expanding CD4 cells came from the host origin because donor-type chimerism never exceeded 0.1%, whereas the total increase of CD4 cells was 10% after the boosts. It is feasible that these cells (possibly T-cell memory cells) were created prenatally, with their rapid proliferation upon postnatal reexposure to the donor antigens in MLR or after the boosts. Allograft rejection by CD4 but not CD8 cells was previously demonstrated by Martin who reported that CD4 cells rejected allogeneic bone marrow cells via a veto mechanism [36]. Veto cells are usually restricted to the CD8 cells and are able to inactivate or delete T cells when they are attacked by these cells [37,38]. It is our hypothesis that some host-type CD4 cells might have acquired memory function during the prenatal exposure to the allogeneic cells and upon postnatal reexposure to these cells became activated and rejected the graft. It has been previously shown that memory T cells are present at low levels in the peripheral blood and can become fully activated after reexposure to the same antigen [39]. Further studies will be done in order to identify the phenotype of the expanding CD4 cells and their role in the allograft rejection. Transplanted mice showed a decrease in anti-donor CTL responses when compared to the control mice. This may be due to the reduced activity of CD8 cells, or to the high production of IL-2 (Fig. 4). Interleukin-2, known as a T-cell growth factor, can promote clonal T-cell expansion during immune responses, but occasionally it can also induce cell apoptosis [40]. Furthermore, it has been previously demonstrated that CD4 T cells stimulated in vitro with recombinant IL-2 occasionally reduce, rather than enhance, the generation of specific CTL activity [41]. Hence, it is feasible to

623

conclude that IL-2, which was released by the expanding host-type CD4 cells, caused a decrease in anti-donor CTL activity observed in our transplanted mice [42,43]. There is increasing evidence that fetuses can generate immune responses and the dogma of prenatal tolerance induction is increasingly challenged [1,2,18–24]. Indeed, it is known that mature T cells can be detected in the human fetal liver by 9 weeks of gestation and blood-borne stem cells can enter developing fetuses at approximately 7 weeks of gestation [16]. It is also known that thymic colonization of early T cells occurs in humans in the first trimester [32], which may explain why prenatal transplants performed in the second trimester have been unsuccessful. There have been no clinical successes with prenatal transplantation in human fetuses with diseases other than SCID. Before further prenatal transplants on non-SCID fetuses can be performed, specific conditions for the induction of a high degree of tolerance must be developed. Understanding of these mechanisms will allow the creation of successful prenatal transplant protocols for congenital disorders and avoidance of complications such as rejection of postnatal transplants and development of late autoimmunity.

Acknowledgments This study was supported by grants from the National Institutes of Health (#K08-HL03603 and #5R01-DK57524). We thank Dr. Maurizio Zanetti for critical review of this manuscript; Dr. Michael Croft from the Institute for Allergy and Immunology for ELISA assay; Biostatistics Shared Resources at the University of California, San Diego for statistical analysis; and Barbara Vickers and Melinda Richards for excellent administrative support.

References 1. Carrier E, Gilpin E, Lee TH, et al. (2000) Microchimerism does not induce tolerance after in utero transplantation and may lead to the development of alloreactivity. J Lab Clin Med 136:224 2. Donahue J, Gilpin E, Lee TH, et al. (2001) Microchimerism does not induce tolerance and sustains immunity after in utero transplantation. Transplantation 71:359 3. Carrier E, Lee TH, Bush MP, et al. (1995) Induction of tolerance in nondefective mice after in utero transplantation of major histocompatibility complex–mismatched hematopoietic cells. Blood 86:4681 4. Carrier E, Lee T-H, Bush MP, Cowan MJ (1997) Recruitment of engrafted donor cells postnatally into the blood with cytokines after in utero transplantation in mice. Transplantation 64(4):627[/1a] 5. Yuh D, Gandy K, Hoyt G, et al. (1996) Tolerance to cardiac allograft induced in utero with fetal liver cells. Circulation 94:304 6. Tanaka SA, Hiramatsu T, Oshitomi T, et al. (1998) Induction of donorspecific tolerance to cardiac xenograft in utero. J Heart Lung Transplant 17:888 7. Flake AW, Roncarolo MG, Puck JM, et al. (1996) Treatment of Xlinked severe combined immunodeficiency by in utero transplantation of paternal bone marrow. N Engl J Med 335:1806 8. Wengler G, Lanfranchi A, Fruscat T, et al. (1996) In utero transplantation of parental CD34 hematopoietic progenitor cells in patient with Xlinked severe combined immunodeficiency (SCID XI). Lancet 348: 1484

624

H. Sefrioui et al./Experimental Hematology 30 (2002) 617–624

9. Touraine JL (1996) In utero transplantation of fetal liver stem cells into human fetuses. J Hematol 5:195 10. Touraine JL (1996) Treatment of human fetuses and induction of immunological tolerance in human by in utero transplantation of stem cells into fetal recipients. Acta Haematol 96:115 11. Flake AW, Zanjani ED (1998) In utero transplantation for thalassemia. Ann N Y Acad Sci 850:300 12. Westgren M, Ringden O, Eik-Nes S, et al. (1996) Lack of evidence of permanent engraftment after in utero fetal stem cell transplantation in congenital hemoglobinopathies. Transplantation 61:1176 13. Billingham RE, Brent L, Medawar PB (1956) Quantitative studies on tissue transplantation immunity. Actively acquired tolerance. Philos Trans R Soc Lond B Biol Sci 239:357 14. Renda MC, Fecarotta E, Dieli F, et al. (2000) Evidence of alloreactive T lymphocytes in fetal liver: implication for fetal hematopoietic stem cell transplantation. Bone Marrow Transplant 25:135 15. Barbey C, Irion O, Helg C, et al. (1998) Characterization of the cytotoxic alloresponses of cord blood. Bone Marrow Transplant 22:26 16. Muench MO, Rae J, Barcena A, et al. (2001) Transplantation of a fetus with paternal Thy-1CD34 cells for chronic granulomatous disease. Bone Marrow Transplant 27:355 17. Pahal GS, Jauniaux E, Kinnon C, et al. (2000) Normal development of human fetal hematopoiesis between eight and seventeen weeks gestation. Am J Obstet Gynecol 183:1029 18. Renda M, Fecarotta E, Maggio A (2000) In utero fetal liver hematopoietic stem cell transplantation: is there a role for alloreactive T lymphocytes? Blood 96:1608 19. Orlandi F, Giambona A, Messana M, et al. (1996) Evidence of induced non-tolerance in HLA-identical twins with hemoglobinopathy after in utero fetal transplantation. Bone Marrow Transplant 18:637 20. Munn D, Zhou M, Attwood J, et al. (1998) Prevention of allogeneic fetal rejection by tryptophan catabolism. Science 281:1191 21. Watts AM, Stanly JR, Shearer MH, et al. (1999) Fetal immunization of baboons induces a fetal-specific antibody response. Nat Med 5:427 22. Gill T, Repetti C, Metlay L, et al. (1983) Transplacental immunization of the human fetus to tetanus by immunization of the mother. J Clin Invest 72:987 23. Shen Z, Mohiuddin M, DiSesa V (1996) Fetal inoculation with donor cells in cardiac xenotransplantation. Ann Thorac Surg 62:1360 24. Moretta A, Locatelli F, Mingrat G, et al. (1999) Characterization of CTL directed towards noninherited maternal alloantigen in human cord blood. Bone Marrow Transplant 24:1161 25. Lu J, Patrene KD, Appasamy PM, et al. (1999) Characteristic of the stage in natural killer cell development in 14.5-day mouse fetal liver using adult bone marrow stroma. Exp Hematol 27:1046 26. Manoussaka MS, Smith RJ, Conlin V, et al. (1997) Fetal mouse NK cell clones are deficient in ly49 expression, share a common broad

27.

28. 29.

30.

31. 32. 33.

34.

35. 36.

37.

38.

39. 40. 41. 42.

43.

lytic specificity, and undergo continuous and extensive diversification in vitro. J Immunol 160:2197 Guller S, LaChapelle L (1999) The role of placental Fas ligand in maintaining immune privilege at maternal-fetal interfaces. Semin Reprod Encrinol 17:39 Flake AW, Zanjani ED (1997) In utero hematopoietic stem cell transplantation. A status report. JAMA 278:932 Flake AW, Roncarolo MG, Puck JM, et al. (1996) Treatment of X-linked severe combined deficiency by the in utero transplantation of CD34 enriched paternal bone marrow. N Engl J Med 335:1806 Andreani M, Manna M, Lucarelli U, et al. (1996) Persistence of mixed chimerism in patients transplanted for the treatment of thalassemia. Blood 87:3494 Harrison MR, Golbus MS, Filly RA (1990) The Unborn Patient. Philadelphia: Saunders Sharma S, Pati HP, Ahuja RK, et al. (1997) Haemopoietic cell composition of human fetal liver, spleen, and thymus. Med Oncol 14:99 Morohashi T, Ogasawara K, Kitaichi N, et al. (2000) Significant MLR but not CTL responses against recipient antigens generated in T cells from bone marrow chimeras recovered from acute GVHD. Bone Marrow Transplant 26:1069 Murray AM, Simm B, Beagley KW (1998) Cytokines gene expression in murine fetal intestine: potential for extratymic T-cell development. Cytokines 10:337 Tretiakova AP, Little CS, Blank KJ, et al. (2000) Rational design of cytotoxic T-cell inhibitors. Nat Biotechnol 18:984 Martin PJ (1996) Prevention of allogeneic marrow graft rejection by donor T cells that do not recognize recipient alloantigen: potential role of a veto mechanism. Blood 88:962 Tscherning T, Claeson MH (1991) Veto-like downregulation of T helper cell reactivity in vivo by injection of semi-allogeneic spleen cells. Immunol Lett 29:223 Martin PJ (1993) Donor CD8 cells prevent allogeneic marrow graft rejection in mice: potential implications for marrow transplantation in humans. J Exp Med 178:703 Springer TA (1994). Traffic signals for lymphocyte recirculation and leukocyte emigration: the multistem paradigm. Cell 76:301 Nelson BH, Willerford DM (1998) Biology of interleukin-2 receptor. Adv Immunol 70:1 Cassell D, Forman J (1991) Two roles for CD4 cells in the control of the generation of cytotoxic T lymphocytes. J Immunol 146:3 Rowley DA, Becken ET, Stach RM (1995) Autoantibodies produced spontaneously by young 1pr-mice carry transforming growth factor  and suppress cytotoxic T lymphocyte responses. J Exp Med 181: 1875 Chavin KD, Qin L, Yon R, et al. (1994) Anti-CD2 mAbs suppress cytotoxic lymphocyte activity by the generation of Th2 suppressor cells and receptor blockade. J Immunol 152:3729