Graft-versus-host disease (GVHD) is a multistep disease process following allogeneic bone marrow transplantation (BMT). It has been postulated that the induction of acute GVHD requires the presence of Peyer patches (PPs). A new tumor necrosis factor (TNF)–deficient strain has been developed that totally lacks PPs and displays the defects characteristic of TNF ablation but not lymphotoxin-associated defects characterized by lack of both PPs and lymph nodes. To determine the necessity of PPs in acute lethal GVHD induction, we transplanted full major histocompatibility complex (MHC)–mismatched grafts into myeloablated TNF knockout recipients. No differences in the survival or GVHD-associated histopathologic lesions were observed between the recipients. We conclude that neither PPs nor host TNF-α is required for the development of acute lethal GVHD in mice that undergo myeloablative conditioning and allogeneic BMT.


Gut-associated lymphoid tissue (GALT) is found throughout the intestine. It consists of the lamina propria of the submucosa, gut cryptopatches,1 intraepithelial lymphocytes (IELs), and the nodule-type tissues similar to a lymph node (LN) in function called Peyer patches (PPs). Development of PPs, cecal patches (CPs), and LNs depends on the expression of certain members of the tumor necrosis factor (TNF) ligand and receptor superfamilies. LNs, CPs, and PPs are absent in lymphotoxin-deficient mice.2 However, the presence of PPs varies between different independently generated TNF and TNF receptor 1 (TNFR1) knockout strains, ranging from nonexistent to virtually normal.3-6

The availability of TNF knockout strains of mice with or without PPs allowed us to address the role of both TNF and PPs in resistance to graft-versus-host disease (GVHD) caused by major histocompatibility (MHC)–mismatched hematopoietic grafts. Previous reports have shown that although TNF-α is produced by diverse types of activated cells,7 only donor-derived TNF is important in the induction of acute lethal GVHD8 as well as leukocyte movement in autoimmune disease.9 PPs are reported to be important for the homing and priming of T-cell effector cells in intestinal GVHD10-12 and have been reported to be essential for GVHD induction in a murine model that does not use conditioning of the host prior to adoptive transfer of the allogeneic donor cells.12 However, as a clinical entity, acute GVHD is encountered primarily in patients who receive a conditioning regimen in preparation for the hematopoietic-cell transplant.

We report here that the absence of TNF and PPs still allowed for acute GVHD induction in models in which myeloablative conditioning was applied before bone marrow transplantation (BMT).

Study design


BALB/c were purchased from the Animal Production Area (National Cancer Institute [NCI] at Frederick, Frederick, MD). C57BL/6 TNFΔ/Δ mice were generated by LoxP/Cre technology.13 Details of the targeting strategy and animal phenotype are reported elsewhere.14,15 C57BL/6 TNF-/- mice were provided by Dr M. Marino (Memorial Sloan-Kettering Cancer Center, New York, NY).16 Normal littermates or C57BL/6 mice purchased from the Animal Production Area were the source of C57BL/6 TNF+/+ mice. Mice were 2 to 4 months of age at the initiation of experiments. Experimental groups were balanced for age and sex of mice. Animals were maintained under specific-pathogen conditions at both NCI-Frederick and the University of Nevada, Reno, facilities. Animal care was provided in accordance with the procedures outlined by the NIH.17 Animal studies were performed at each of the 2 animal facilities according to approved protocols and in accordance with the Animal Care and Use Committees of each institution.

Figure 1.

Acute GVHD induced after allogeneic BMT using extensive conditioning in mice with or without PPs. (A) TNFΔ/Δ and TNF-/- mice develop mesenteric LNs and cecal patches (CPs). TNF-/- mice develop PPs, but TNFΔ/Δ mice do not. Not all PPs formed in the TNF-/- mice are visible. Fresh mouse organs, from representative adult naive animals, were photographed using Nikon Coolpix 2500 digital camera set to macro mode. (B) TNFΔ/Δ, TNF-/- mice, or WT mice were used as recipients of allogeneic BM grafts in combination with 3 × 107 spleen cells (SCs) as a source of T cells (n = 9-11 mice/group/experiment). Some groups received BM only to control for non–GVHD-associated changes (n = 3-8 mice/group/experiment). Combined results of 2 independent experiments are shown. No significant differences were observed in mice that received BM and spleen cells. (C) Kinetics of weight loss in one of the 2 independent experiments represented in panel B. (D) TNFΔ/Δ or WT mice were used as recipients of allogeneic BM grafts in combination with 2.5 or 1.25 × 107 spleen cells (SCs) as a source of T cells (n = 5-8 mice/group). Some groups received BM only to control for non–GVHD-associated changes. Representative results of 2 independent experiments are shown for survival kinetics. No significant differences were observed between experimental groups for either dose of spleen cells.

Induction of GVHD

To induce acute GVHD associated with allogeneic BMT, recipient C57BL/6 (H2b) TNF+/+, TNFΔ/Δ, or TNF-/- mice received myeloablative total body irradiation (950 cGy) from a 137Cs source followed by intravenous infusion of 107 bone marrow cells (BMCs) and splenocytes from BALB/c (H2d) donor mice. Some groups did not receive splenocytes to monitor non–GVHD-associated morbidity. Recipient mice were given amoxicillin ad lib in their drinking water for 2 weeks beginning 7 days before transplantation. Mice were monitored and weighed weekly. All moribund mice were humanely killed. Survival data were plotted by the Kaplan-Meier method and analyzed by the log-rank test.


Formalin-fixed, paraffin-embedded tissue sections were stained with hematoxylin and eosin and evaluated and graded in coded fashion by a veterinary pathologist as previously described.18 A semiquantitative scale for histopathologic changes ranked from 0 to 4 was used. Cumulative histopathology scores were calculated based on the sum of individual scores for 3 to 5 parameters in each organ (villous blunting, crypt-cell hyperplasia, crypt-cell apoptosis, and epithelial-cell sloughing in the small intestine; goblet-cell depletion, inflammation, sloughing of epithelial cells into the lumen, crypt-cell apoptosis, and hyperplasia in the colon; and vacuolation, necrosis, and inflammation in the liver). Comparison of cumulative histopathologic scores were analyzed the Kruskal-Wallis and Dunn multiple comparison tests (P < .05). Images were visualized using an Olympus Vanox AHBS3 microscope with an Olympus SPlan Apo 20×/0.70 numeric aperture objective (Olympus, Woodbury, NY). A Diagnostic Spot RT color digital camera using Spot software version 4.0.2 was used to acquire the images (Diagnostic Instruments, Sterling Heights, MI).

Results and discussion

Absence of PPs in TNFΔ/Δ mice

TNFΔ/Δ mice completely lack PPs,15 recapitulating the previously reported phenotype of TNFR1 knockout.4 However, apart from the complete absence of PPs, TNFΔ/Δ mice shared characteristic phenotypic features with previously reported TNF-/- mice.3,6,16 In particular, both TNF-/- and TNFΔ/Δ mice develop a CP and all LNs, including mesenteric nodes (Figure 1A). In contrast, TNF-/- mice present with PPs, whereas the TNFΔ/Δ mice completely lack PPs. We therefore compared TNF-/- and TNFΔ/Δ mice in an experimental GVHD model in which myeloablative conditioning was applied.

The absence of PPs does not confer protection from acute lethal GVHD using models with cytoreductive conditioning and hematopoietic-cell rescue. Using normal littermates (wild-type [WT]), TNFΔ/Δ, and TNF-/- mice as recipients of fully allogeneic BM- and splenic-cell grafts allowed us to directly test the necessity of PPs in GVHD induction. As demonstrated, TNFΔ/Δ, TNF-/-, and WT recipients showed similar kinetics of mortality (Figure 1B) and weight loss (Figure 1C) from acute GVHD following allogeneic BMT. In addition, no statistical differences in GVHD mortality between TNFΔ/Δ and WT recipients were observed when lower doses of spleen cells were used to induce GVHD (Figure 1D).

Figure 2.

No difference in intestinal GVHD in TNF-deficient hosts using extensive conditioning and allogeneic BMT at either early or late time points. (A) Photomicrograph of colonic tissue sections taken from animals after GVHD induction. Moribund mice from the experiment presented in Figure 1C were humanely killed at day 25 after BMT. Colonic tissue was sectioned and stained by hematoxylin and eosin. Representative photomicrographs are depicted. White arrows designate areas of inflammation. Black arrows designate sloughing of epithelium. (B) No difference in GVHD-associated histologic changes observed in the small intestine, large intestine, and liver of TNFΔ/Δ or WT recipients of allogeneic BM grafts in combination with 3 × 107 spleen cells (SCs) (n = 3/group). Some groups received BM only to control for non–GVHD-associated changes. Mice were assessed 6 days after transplantation. Tissues were assessed for histologic changes in 3 to 4 parameters as described in “Study design” and the sums of these scores are represented for each tissue.

Intestinal GVHD in mice with or without PPs following cytoreductive conditioning and allogeneic BMT

Histopathologic analysis of the liver and intestine from moribund animals revealed no differences in organ-specific GVHD-associated pathologic lesions in the gastrointestinal tract in WT, TNFΔ/Δ, or TNF-/- moribund animals 25 days after BMT (Figure 2A). As summarized in Figure 2B, TNFΔ/Δ and WT recipients were also evaluated at day 6 after BMT for histopathologic lesions associated with GVHD. GVHD-associated changes were evident, particularly in the colon and small intestines. However, there were no significant differences in the cumulative histopathologic scores in TNFΔ/Δ and WT recipient mice of allogeneic BM and spleen cells (Figure 2B). Specifically, no differences in the frequency or grade of crypt hyperplasia and apoptosis were observed in the small and large intestines of TNFΔ/Δ, TNF-/-, and WT recipients nor were there differences in goblet-cell depletion or the presence of sloughed epithelial cells in the crypt lumen at day 6 or day 28 after BMT (Figure 2A-B). Similar frequency and grade of subacute inflammation and hepatocellular vacuolation were observed in the livers of all recipient types (Figure 2B and data not shown).

Therefore, the data demonstrate that in a BMT model using myeloablative conditioning and hematopoietic-cell rescue, neither host-derived TNF nor PPs were required for induction or progression of acute GVHD. These results are in contrast with a recent report that PPs are required for the induction of acute GVHD.12 That report used administration of anti-interleukin 7 receptor (anti-IL7R) antibodies to prevent PP formation that may have altered other tissues in the recipient that could affect GVHD progression. Importantly, the study used an adoptive lymphocyte transfer model that involved no or only myelosuppressive conditioning of the recipient and lacked hematopoietic rescue. The use of myeloablative conditioning and allogeneic BM reconstitution is more reflective of clinical BMT regimens and most likely results in the release of inflammatory cytokines that can alter the requirements for GVHD induction and pathobiology. We and others have recently found that the presence or absence of cytoreductive conditioning of the host also results in markedly contrasting effects on the role of the chemokine receptor, CCR5, in GVHD pathobiology.19,20 Wysocki and colleagues have shown that expression of proinflammatory chemokines important in the trafficking of alloreactive T cells into target tissues was dependent on the use of a myeloablative conditioning regimen.20 Therefore, these data demonstrate that the use of myeloablative conditioning has a marked impact on acute GVHD pathobiology and that PPs are not necessary for its induction or progression when such conditioning is applied.


We thank Dr M. Anver (NCI-FCRDC Pathology/Histotechnology Laboratory, Frederick, MD) for invaluable help with pathology. We also thank Steve Stull, Terri Stull, and Melinda Berthold for assistance. The contents of this publication do not necessarily reflect the views or policies of the Department of Health and Human Services, nor does mention of trade names, commercial products, or organizations imply endorsement by the US government.


  • Reprints:
    Lisbeth A. Welniak, Department of Microbiology and Immunology, University of Nevada School of Medicine, Mail Stop 199, Reno, NV 89557; e-mail: lwelniak{at}
  • Prepublished online as Blood First Edition Paper, September 13, 2005; DOI 10.1182/blood-2004-11-4565.

  • Supported in whole or in part with US Federal funds from the National Cancer Institute (NCI), National Institutes of Health (NIH), under contracts NO1-CO-12400 and RO1 CA102282 (W.J.M.), R01 AI 34495, 2R37 HL56067, and R01 HL63452 (B.R.B.) and P20 RR016464 (L.A.W.) and in part by a grant from the University of Nevada Junior Research Grant Fund (L.A.W.).

  • The publisher or recipient acknowledges right of the US government to retain a nonexclusive, royalty-free license in and to any copyright covering the article.

  • The publication costs of this article were defrayed in part by page charge payment. Therefore, and solely to indicate this fact, this article is hereby marked “advertisement” in accordance with 18 U.S.C. section 1734.

  • Submitted December 8, 2004.
  • Accepted August 12, 2005.


View Abstract