American Joumnal of Patbology, Vol. 140, No. 5, May 1992Copyight © American Association of Pathologists

Hemagglutination and Graft-versus-HostDisease in the Severe CombinedImmunodeficiency MouseLymphoproliferative Disease Model

Samuel J. Pirruccello, Hirokazu Nakamine,Kirk W. Beisel, Kimberly L. Kleveland,Motohiko Okano, Yuichi Taguchi,Jack R. Davis, Mark L. Mahloch, andDavid T. PurtiloFrom the Department ofPathology and Microbiology,University ofNebraska Medical Center, Omaha, Nebraska

In the course of evaluating the severe combined im-munodeficiency mouse-human peripheral bloodlymphocyte (SCID-PBL) model of lymphoprolifera-tive disease, we noted hemagglutination occurring inperipheral blood smears of mice with serum humanimmunoglobulin levels greater than 1.0 mg/ml. Thehemagglutinating process was mediated by humananti-mouse red cell antibodies of the IgM class,peaked atfive to seven weekspost-transfer of5 to 7 X107 human PBL and was generally self limiting.However, death resulted in some mice when serumimmunoglobulin levels were greater than 3.0 mg/ml.The most severely affected mice had hemagglutina-tion induced congestion of liver, lungs and spleenSeveral mice also had lesions consistent with graft-versus-host disease (GVHD) including focal hepaticnecrosis and destruction of mouse splenic hemato-poietic elements. The lesions associated with hemag-glutination and GVHD in SCID-PBL mice are distinctfrom those associated with EBV-induced lymphopro-liferation Recognition of these pathologic processesare required for a thorough understanding of theSCID-PBL model. (Am J Pathol 1992, 140:1187-1194)

mice are being extensively utilized as models to study thein vivo biology of normal, virus infected and malignanthuman cells.12-20

Short term engraftment of mature human T and B lym-phocytes can be demonstrated in the SCID mouse fol-lowing intraperitoneal transfer of human peripheral bloodmononuclear cells from EBV seronegative donors.13,14When peripheral blood lymphocytes (PBL) are trans-ferred from Epstein-Barr virus (EBV) seropositive humandonors, EBV-induced lymphoproliferative disease occursin the SCID mice. Although originally proposed as amodel of EBV-induced lymphoma, we and others havedemonstrated that this model of lymphomagenesis moreclosely resembles EBV-induced lymphoproliferation thatfrequently occurs in immunocompromised patients in theabsence of specific cytogenetic changes.21-24

Although graft-versus-host disease (GVHD) would beexpected to occur following xenografting in the SCID-PBL model, Mosier et al13,14 and Cannon et a123 reportedthat GVHD was minimal or nonexistent. In contrast,Bankert et a125 reported that GVHD-related processeshad occurred in this model system and were likely re-sponsible for the inability of the SCID-PBL mice to gen-erate primary antibody responses to protein antigens.They also reported that the SCID-PBL mice had humanIgM bound to the surface of erythrocytes as demon-strated by immunofluorescence or agglutination with anti-human IgM.25 In the course of evaluating the lymphopro-liferative disease model we also noted hemagglutinationand GVHD-related lesions in a high percentage of SCID-PBL mice engrafted with PBL from normal human do-nors.22 The extent of the hemagglutinating disease wasdirectly related to the level of human immunoglobulin in

The CB.1 7 scid/scid (SCID) mouse, originally describedby Bosma et al,1 is homozygous for a mutation that dis-allows generation of functionally rearranged immuno-globulin and T cell receptor genes.1- The lack of matureT and B lymphocytes results in a severe combined im-munodeficiency phenotype.911 Because the SCID phe-notype allows engraftment of xenogeneic tissues these

Supported by Public Health Service CA 30196 and CA 36727 from theNational Cancer Institute, Department of Health and Human Services, theState of Nebraska Department of Health, LB506, and the Lymphoprolif-erative Research Fund.

Accepted for publication December 13, 1991.Address reprint requests to Dr. Samuel J. Pirruccello, Department of

Pathology and Microbiology, University of Nebraska Medical Center, 600South 42nd Street, Omaha, NE 68198-6495.

1187

1188 Pirruccello et alAJP May 1992, Vol. 140, No. 5

the SCID-PBL mouse serum. We report here the charac-teristics and lesions associated with hemagglutinationand GVHD in the SCID-PBL model.

Materials and Methods

Donors

After receiving informed consent, approximately 200 mlof heparinized blood was obtained from normal, EBV se-ropositive or seronegative human donors. All normal do-nors were healthy laboratory workers who were not in-volved with this project and therefore did not handle theSCID-PBL mice, cells, or tissues. These precautions weretaken to circumvent accidental inoculation of autologousEBV-infected lymphoid cells into laboratory staff. Periph-eral blood mononuclear cells (PBL) were isolated byFicoll-hypaque density gradient centrifugation, washedwith phosphate buffered saline (PBS) and counted. Cellswere resuspended in sterile PBS at 5-7 x 107 cells/mlprior to transfer to the SCID mice. All procedures werecarried out with approval of the Institutional Review Boardof the University of Nebraska Medical Center (IRB# 270-89).

Mice

C.B-1 7 SCID breeding trios were originally obtained fromMcLaughlin Research Institute (Great Falls, MT). Animalsutilized for these studies were bred and maintained in apathogen free environment within a Class-Il laminar flowsafety hood (Baker, Stanford, ME) in micro-isolator cagesat the Animal Resource Facility, University of NebraskaMedical Center. All mice were screened for production ofendogenous immunoglobulin (i.e., "leaky" phenotype) byassaying for serum immunoglobulin by radial immunod-iffusion (Chemicon, El Segundo CA) at 6 weeks of age.Mice without detectable serum immunoglobulin were uti-lized as recipients. As prophylaxis against Pneumocystiscarinii pneumonia, mice were maintained on tri-methoprim-sulfamethoxazole (0.32 g/ml trimethoprimand 1.6 g/ml sulfamethoxazole) (Biocraft, Elmwood, NJ)in their dnnking water for three of every seven days. Theremainder of the week the mice were provided with ster-ile, acidified water.

PBL Transfer, Monitoring, and TissueProcessing

Mice were inoculated with 5-7 x 107 PBL by intraperito-neal injection. Twenty eight of the 32 mice were engraftedwith PBL from EBV seronegative donors and were sec-ondarily inoculated with B95-8 EBV at six weeks post

transfer of human PBL as described previously.21 22 Theremaining 4 mice were engrafted with PBL from EBV se-ropositive donors. At 3 weeks post transfer of PBL andweekly thereafter approximately 0.1 to 0.2 ml of bloodwas obtained from each mouse by retro-orbital sinusplexus puncture. Serum was assayed by radial immun-odiffusion for human immunoglobulin to assess engraft-ment and peripheral blood smears were made and eval-uated for hemagglutination. The degree of hemaggluti-nation was retrospectively assigned a relative value of 1to 3 + by the following criteria: negative, no evidence ofhemagglutination; 1 +, mild hemagglutination with occa-sional small aggregates of 3-4 red cells each; 2+,marked hemagglutination with multiple small and large,multicellular aggregates; and 3+, characterized as 2 +hemagglutination or severe anemia with physical signs ofdisease.

The mice were monitored three times weekly forchanges in physical appearance such as respiratory dis-tress, organomegaly, ruffled fur, decreased mobility, in-ability to eat, diarrhea, or weight loss. Mice with evidenceof hemagglutination on peripheral smear and physicalsigns of distress were euthanized and autopsied. At nec-ropsy the mice were examined for gross evidence of or-ganomegaly, tissue necrosis or lymphoproliferative le-sions. The liver, spleen, heart, lungs, and kidneys as wellas grossly visible lymphoproliferative lesions were re-moved, fixed in 10% buffered formalin and processed forroutine tissue sections and microscopic evaluation. Thetissues were dehydrated and embedded in paraffin, cutat 5 ,um and stained with hematoxylin and eosin. All pro-tocols were carried out in accordance with the guidelinesof the Animal Review Committee, University of NebraskaMedical Center (ARC# 89-104-02).

Immunohistochemical AnalysisImmunohistochemical examination was performed onblocks of tissues which were snap frozen in liquid nitro-gen or fixed in formalin. The latter were stained with an-tibodies to a T cell related antigen CD45RO (UCHL-1), aB cell related antigen CD20 (L26), and HLA-DR (LN3) bythe avidin-biotin-peroxidase complex (ABC) method.3233Background staining was determined by substituting pu-rified mouse immunoglobulin as primary antibody fol-lowed by ABC (negative control). Biotinylated secondaryantibodies and ABC were purchased from Vector (Burl-ingame, CA). Following reaction with 3,3'-diamino-benzidine the sections were counterstained with hema-toxylin and examined with the light microscope.

Flow Cytometric Analysis

Erythrocytes obtained from mice with 3 + hemagglu-tination were evaluated by flow cytometric analysis for

GVHD in SCID-PBL Mice 1189AJP May 1992, Vol. 140, No. 5

class of surface bound human immunoglobulin by mod-ification of previously described techniques.26 Approxi-mately 0.2 ml peripheral blood collected in sodium hep-arin was diluted in PBS with 5% fetal bovine serum(Gibco, Grand Island, NY) and .01% sodium azide to aconcentration of 106 erythrocytes/ 0.1 ml and aliquoted at106 cells/tube. Individual red cell aliquots were incubatedfor 30 minutes at 40C with saturating concentrations offluorescein (FITC) conjugated goat anti-human IgG, IgMand IgA heavy chains and kappa and lambda lightchains (Tago, Burlingame, CA). Background immunoflu-orescence was determined by staining an aliquot of cellswith FITC conjugated goat anti-mouse IgG. Cells wereanalyzed on a three decade log fluorescence scale forpercent positivity and mean channel fluorescence follow-ing background subtraction. All assays were performedon a Coulter EPICS C flow cytometer (Coulter, Hialeah,FL) using Coulter software.

Results

Hemagglutination

Of 32 mice receiving 5-7 x 107 human PBL by intraper-itoneal injection, 23/32 (71 %) showed evidence of hem-agglutination on peripheral smear (Fig. 1A-C). The de-gree of hemagglutination was variable from mouse tox ^ _0~~~~~~~~~~~~rIV:*

~ v

*

mouse at the same initial cell dose of PBL but was directlyrelated to the concentration of human immunoglobulin inthe serum (Fig. 2). Nine of 32 mice showed no evidenceof hemagglutination in peripheral blood smears and allhad human immunoglobulin levels less than 1.0 mg/ml.Fourteen mice which had either 1 + or 2 + hemaggluti-nation with no physical signs of distress had immuno-globulin levels between 1.0 and 3.0 mg/ml. Nine micewere severely affected with 2 + hemagglutination or se-vere anemia on peripheral smear and physical signs ofdistress (i.e., 3 + hemagglutination). Seven of these ninemice (78%) had immunoglobulin levels greater than 3.0mg/ml.

The hemagglutinating disease process peaked at 5 to7 weeks post-transfer of 5-7 x 107 human PBL and wasgenerally self limiting. The timing and appearance ofmore severe hemagglutination was heralded by earlierdevelopment of measurable serum levels of human im-munoglobulin. Of the 32 mice engrafted, three died of thedisease and six were euthanized with 3 + disease. Phys-ically the mice presented with features of GVHD includ-ing wasting and diarrhea. One of the mice which haddied during the peak of hemagglutination developed cy-anosis and necrosis of the digits.

Gross and Microscopic PathologyGross and microscopic examination was performed onthe six mice euthanized for 3+ disease. Grossly, there

Figure 1. Pbotomicrograpbs ofperipberal blood smears obtainedat 5 to 7 weeks post-transfer of5 x 107 PBL demonstrating bem-agglutination (a,b) and severe anemnia (c). a: 1+ hemagglutina-tion with occasionalfragmented red cells and small red cell ag-gregates (arrow). b: 2 + bemagglutination, numerous large mul-tiellular red cell aggregates arepresent. c: 3 + bemagglutinationcbaracterized by severe anemia with anisocytosis andpolychroma-sia (magnification x 1,000).

1190 Pirruccello et alAJP Alaiv 1992, Vol 140, No. 5

Flow Cytometric Analysis4

P-

E

E

%.

Negative 1 2 . 3

Hemagg lut i nation

egree 0f maggc lunnatisltion onoperipheral s zearat S-'ueek,s postcell transnfer plotted cagafinst hunian imiuinillRloglobulliZlev els inz moiise serumi. 1Hemagglutinaition wcas cAses.sed as de-scribed iider Mlethods Inimunoglobdlin lei!els rflect the gr-ou)ptetaz ialues + onle standaird dev iation. [he negatile grouzp) rep-re.seints the mveano olf three mice who had niea,surchle level.s cfho/-nman imnmnunoglohuzlin. 77)e reiaining sik mice in this grolp didniot develop mnecsurahle imimuniioglobulin levels in the time periodsttudied. Ihe nittszber-ofaninitials in thMe renaining grouips l ere: I +(i = 7), 2 + (i = 7), -±+ (ni = 9).

was mottling of the liver and spleen consistent with con-

gestion in two of the six mice. This finding was confirmedon microscopic examination which demonstratedmarked congestion and hemostasis by hemagglutinationin the liver, spleen, and lungs. The livers of all six mice

contained numerous iron laden macrophages with vari-

able sinusoidal dilatation as a result of hemostasis. Focalareas of necrosis were evident in the livers of three mice(Fig. 3a). Periportal and perivascular cellular infiltrateswere present in the livers of all six mice (Fig. 3b). Theseinfiltrates consisted predominantly of human T cells withsmaller numbers of B cells, plasma cells and mouse neu-

trophils (Figs. 4a-c). Grossly, splenomegaly was evidentin two mice while microscopically all mice showed some

degree of human lymphocyte infiltration and proliferationwithin the spleen. There was a loss of mouse hematopoi-etic elements and fibrosis in the spleens of three of the sixmice (Figs. 3c and 3d). Immunohistochemistry demon-strated mouse stromal elements and human T cells re-

maining in these fibrotic spleens. One mouse had com-

pletely infarcted the spleen just prior to necropsy. Splenicinfarction in this mouse appeared to have resulted frommassive hemagglutination and congestion coupled withhuman lymphocyte infiltration and proliferation. All micehad small, focal aggregates of lymphoid cells with a pre-dominant plasma cell component scattered throughoutthe visceral and parietal peritoneum (Fig. 3e).

Flow cytometric analysis for cell surface bound humanimmunoglobulin was performed on red cells obtainedfrom four mice with 3+ hemagglutinating disease (Fig.5). These mice represented three separate donors in-cluding two EBV seronegatives and one EBV seroposi-tive. After background subtraction, all four specimensdemonstrated bound immunoglobulin to be IgM withvariable amounts of IgG. We had difficulty demonstratinglight chains at the same percent positivity as surfacebound IgM however, those demonstrable were predom-inantly of the kappa class. To determine if circulating,pre-formed anti-mouse red cell antibodies were detect-able in the sera of the three donors, we screened donorserum by flow cytometry on erythrocytes obtained from anaive SCID mouse. In none of the three donors could wedefinitively identify circulating anti-mouse red cell anti-bodies of IgG, IgM or IgA heavy chain class (not shown).

Discussion

In 1954 Barnes and Loutit27 reported a secondary dis-ease occurring in mice which had received an allogeneicspleen graft following lethal irradiation. This secondarydisease (the primary disease being radiation sickness)was characterized by severe wasting, diarrhea, and skinlesions. A few years later Billingham and Brent describeda runting syndrome occurring in newborn mice trans-planted with parent strain spleen cells.28 This runting syn-drome was characterized by severe growth retardation,diarrhea, hyperplasia of the lymphatic system followedby hypoplasia, skin lesions, and focal necrosis of the liver.These phenomena were thought to be the result of animmmunologic attack of the graft on the host. Furthertransplantation studies using parental/F1 mouse systemsclearly established the concept of graft-versus-host dis-ease and that this process was based on a reaction ofengrafted, immunologically competent cells against animmunodeficient host.9 The pathology associated withallogeneic GVHD in humans became well defined withthe advent of bone marrow transplantation therapy.30 Inacute GVHD the hallmark lesion is selective epithelialdamage of target organs including the skin, liver andgastrointestinal tract.

Initial descriptions of the SCID-PBL model13,14 23 re-ported that obvious GVHD did not occur following intra-peritoneal transfer of up to 5 x 107 mature human pe-ripheral blood mononuclear cells. In contrast, Bankert eta125 observed GVHD-related lesions in skin and liver ofthe SCID-PBL mice following intraperitoneal transfer of asfew as 1 x 107 human PBL. GVHD-related lesions werereported to be present at necropsy fourteen weeks post-

-T-

o-

GVHD in SCID-PBL Mice 1191AJP May 1992, Vol. 140, No. 5

3WI ,. .'s.' '.,.*g !af.,

:1' *,,

. .

. .

=:<<||:4^ss EL: -o .

Figure 3. Photomicrograpbs of tizssue sectionsfrom SCID micewith 3 + hemnagglutination demonstrating associated lesions.a: Liver section showingfocal hepatic necrosis (magnificationx200). b: Liver section showing periportal cellular infiltrate(magnification x 400). c: Section of nonengrafted, SCIDmouse control spleen showing splenic bematopoiesis (magni-fication X200). d: Section ofspleen showing fibrosis and lossof mouse bematopoietic elements (magnification X200). e:Kidney section sbowing bumaanplasma cellproliferation in thekidney capsule (magnification x400).

transfer of PBL, concurrently with extensive lymphopro-liferative disease. When 1 x 108 PBL were transferred,the mice exhibited clinical symptoms of GVHD at 3-4weeks post-transplantation and all mice were dead by sixweeks. Intermediate cell doses clearly demonstrated thatthe timing of GVHD was dose related. They also hadnoted that the mouse erythrocytes had human IgMbound to the plasma membranes.

Consistent with Bankert et a125 we have also foundthat significant GVHD occurred in SCID-PBL mice. In thecourse of our evaluation of the model we noted a hemag-glutinating disease process in many animals with evi-dence of B cell engraftment.22 Affected mice all had se-rum immunoglobulin levels greater than 1.0 mg/ml whileunaffected mice had levels less than 1.0 mg/ml. Flowcytometric evaluation of erythrocytes obtained from af-fected mice demonstrated human IgM and small

amounts of human IgG bound to the red cell surface.These results confirm Bankert's earlier observation.25

Coombs positive hemolytic anemia is frequently as-sociated with acute GVHD in allogeneic mouse models-9and has also been documented in humans with acuteGVHD following bone marrow transplantation.31 It seemslikely that in the SCID-PBL xenograft model the anti-mouse red cell response is analogous. Hemagglutinationwas generally self-limited at the cell doses used for thesestudies. Thirteen of twenty two affected mice recoveredand of these, only seven went on to develop lymphopro-liferative disease. This finding indicates that despite thecontinued presence of mouse red cell antigen the spe-cific B cell clone(s) either became tolerant or more likelyattenuated. The fact that mice engrafted with PBL fromEBV seronegative donors show only a low level, transientexpression of serum immunoglobulin is consistent with

1192 Pirruccello et alAJP May 1992, Vol. 140, No. 5

tne eventual loss of these clones.1314'21-23 Given thatwith our protocol hemagglutination was often evidentprior to EBV inoculation also indicates that the anti-redcell response was not EBV driven.

The fact that the anti-red cell immunoglobulin was pre-dominantly IgM is consistent with a primary anti-mouseresponse. It is however, possible, that the specificity ofthese antibodies for mouse red cells represented natu-rally occurring antibodies analogous to anti-blood group

Figure 4. Immunohistologic preparation of SCID mouse liversections demonstrating periportal and perivascular humanT-cell infiltrates a: Section of affected SCID lwer stained wuitnegative control antihody. b,c: Sections of affeted SCID liverstained with anti-CD45RO (UCHL-1) (magniftcation x400).

antigens in humans. We could not however, readily dem-onstrate circulating antibodies in donor serum by flowcytometry. The possibility that donors were previouslysensitized to mouse antigens also seems unlikely sinceonly one of the donors had previously worked with mice.Alternatively, these antibodies may represent silent, po-tentially autoreactive clones (i.e. with primary reactivity tohuman red cells) that are reactivated in vivo as part of theGVHD response. We did not directly test this possibility.

LiIi~ ControlFigure 5. Single parameter flow cytometrichistograms ofred cellsffrom SCID mouse with3+ hemagglutination. Control cells werestained with goat anti-mouse IgG FITC. Testcells were stained with goat anti-human Igspecific for heavy and light chains as indi-cated in the histograms and cells were ana-lyzed on a three decade log fluorescencescale. After background subtraction, IgM =96% positive, IgG = 24% positive, and kappa= 12% positive. IgA and lambda were nega-tive.

LOG GREEN FLUORESCENCE

. _ _ _ _ _ _ _

m

;Dz

H

'4

Lz4

GVHD in SCID-PBL Mice 1193AJP May 1992, Vol. 140, No. 5

The transfer of EBV negative human PBL to SCIDmice results in a peak of lymphocyte engraftment that isdependent on the initial cell dose.13'14'25 Many of themeasures of human lymphocyte engraftment such as ris-ing serum immunoglobulin levels and splenic T cell pro-liferation may really be manifestations of an acute graft-versus-host response. We and others have previouslydocumented GVHD-like lesions in the skin of SCID-PBLmice22 25however, these changes were modest in com-parison to the skin and gut lesions characteristic of GVHDfollowing bone marrow allografting.' It must be kept inmind that the SCID-PBL model system represents xe-nografting of tissues between widely divergent species.For this reason, the more classic GVHD lesions seen inthe gut and skin following allografting may not predomi-nate. The disparity in protein antigens would be expectedto elicit a vigorous antibody response which was seen inthe anti-red cell antibodies. We are not sure whether otherspecificities were elicited since we did not specificallylook for them. The cellular arm of the GVH responsewould be dependent on the ability of human T cells toefficiently traffic to target tissues and to effectively en-gage mouse MHC antigens through the T cell receptor.Since mouse MHC class and class 11 antigens would beexpressed on hematopoietic elements in the SCIDmouse spleen, these cells would represent good cellulartargets. Loss of splenic hematopoiesis and subsequentfibrosis was documented in several mice with the mostsevere hemagglutinating disease. Further, the occur-rence of focal hepatic necrosis originally described inmouse GVHD models28 -9 also appears to be a compo-nent of GVHD in the SCID-PBL model. We and others24have documented these lesions in the absence of obvi-ous LPD indicating a process distinct etiologically fromEBV-induced lymphoproliferation.

Despite the development of hemagglutination andacute GVHD in SCID mice engrafted with human PBLthis model remains very useful for investigating EBV-induced lymphoproliferative disease. At cell doses be-tween 1 x 107 and 5 x 107 human PBL, the majority ofmice will develop lymphoproliferative disease and only asmall fraction of animals will be lost to hemagglutinationor acute GVHD. It is important, however, to distinguishthe lesions associated with these processes. A morethorough understanding of the SCID-PBL model isneeded to assure appropriate utilization and interpreta-tion of experimental results.

References

1. Bosma GC, Custer RP, Bosma MJ: A severe combined im-munodeficiency mutation in the mouse. Nature 1983,301:527-530

2. Hendrickson EA, Schlissel MS, Weaver DT: Wild-type V(D)Jrecombination in scid pre-B cells. Mol Cell Biol 1990,10:5397-5407

3. Kim MG, Schuler W, Bosma MJ, Marcu KB: Abnormal re-combination of Igh D and J gene segments in transformedpre-B cells of scid mice. J Immunol 1988,141:1341-1347

4. Malynn BA, Blackwell TK, Fulop GM, Rathbun GA, FurleyAJW, Ferrier P, Heinke BL, Phillips RA, Yancopoulos GD, AltFW: The scid defect affects the final step of the immuno-globulin VDJ recombinase mechanism. Cell 1988, 54:453-460

5. Carroll AM, Hardy RR, Petrini J, Bosma MJ: T cell leakinessin scid mice. Curr Top Microbiol Immunol 1989, 152:117-123

6. Fulop GM, Phillips RA: The scid mutation in mice causes ageneral defect in DNA repair. Nature 1990, 347:479-482

7. Schuler W, Weiler IJ, Schuler A, Phillips RA, Rosenberg N,Mak TW, Keamey JF, Perry RP, Bosma MJ: Rearrangementof antigen receptor genes is defective in mice with severecombined immune deficiency. Cell 1986, 46:963-972

8. Petrini JH, Carroll AM, Bosma MJ: T-cell receptor gene re-arrangements in functional T-cell clones from severe com-bined immune deficient (scid) mice: reversion of the scidphenotype in individual lymphocyte progenitors. Proc NatlAcad Sci USA 1990, 87:3450-3453

9. Custer RP, Bosma GC, Bosma MJ: Severe combined im-munodeficiency (SCID) in the mouse. Pathology, reconsti-tution, neoplasms. Am J Pathol 1985, 120:464-477

10. Dorshkind K, Keller GM, Phillips RA, Miller RG, Bosma GC,O'Toole M, Bosma MJ: Functional status of cells from lym-phoid and myeloid tissues in mice with severe combinedimmunodeficiency disease. J Immunol 1984, 132:1804-1808

11. Dorshkind K, Pollack SB, Bosma MJ, Phillips RA: Naturalkiller (NK) cells are present in mice with severe combinedimmunodeficiency (scid). J Immunol 1985,134:3798-3801

12. Namikawa R, Kaneshima H, Lieberman M, Weissman IL,McCune JM: Infection of the SCID-hu mouse by HIV-1. Sci-ence 1988, 242:1684-1686

13. Mosier DE, Gulizia RJ, Baird SM, Wilson DB: Transfer of afunctional human immune system to mice with severe com-bined immunodeficiency. Nature 1988, 335:256-259

14. Mosier DE, Gulizia RJ, Baird SM, Spector SM, Spector D,Kipps TJ, Fox RI, Carson DA, Cooper N, Richman DD, Wil-son DB: Studies of HIV infection and the development ofEpstein-Barr virus-related B cell lymphomas following trans-fer of human lymphocytes to mice with severe combinedimmunodeficiency. Curr Top Microbiol Immunol 1989,152:195-199

15. McCune JM, Namikawa R, Shih CC, Rabin L, Kaneshima H:Suppression of HIV infection in AZT-treated SCID-hu mice.Science 1990, 247:564-566

16. McCune JM, Namikawa R, Kaneshima H, Shultz LD, Lieber-man M, Weissman IL: The SCID-hu mouse: murine modelfor the analysis of human hematolymphoid differentiationand function. Science 1988, 241:1632-1639

17. Ghetie MA, Richardson J, TuckerT, Jones D, UhrJW, Vitetta

1194 Pirruccello et alAJP May 1992, Vol. 140, No. 5

ES: Disseminated or localized growth of a human B-cell tu-mor (Daudi) in SCID mice. Int J Cancer 1990, 45:481-485

18. Duchosal MA, McConahey PJ, Robinson CA, Dixon FJ:Transfer of human systemic lupus erythematosus in severecombined immunodeficient (SCID) mice. J Exp Med 1990,172:985-988

19. Kamel-Reid S, Letarte M, Sirard C, Doedens M, GrunbergerT, Fulop G, Freedman MH, Phillips RA, Dick JE: A model ofhuman acute lymphoblastic leukemia in immune-deficientSCID mice. Science 1989, 246:1597-1600

20. Mosier DE, Gulizia RJ, Baird SM, Wilson DB, Spector DH,Spector SA: Human immunodeficiency virus infection of hu-man-PBL-SCID mice. Nature 1991, 251:791-794

21. Okano M, Taguchi Y, Nakamine H, Pirruccello SJ, Davis JR,Beisel KW, Kleveland KL, Sanger WG, Fordyce RR, PurtiloDT: Characterization of Epstein-Barr virus-induced lym-phoproliferation derived from human peripheral bloodmononuclear cells transferred to severe combined immun-odeficient mice. Am J Pathol 1990,137:517-522

22. Purtilo DT, Falk K, Pirruccello SJ, Nakamine H, Kleveland K,Davis JR, Okano M, Taguchi Y, Sanger WG, KW Beisel:SCID mouse model of Epstein-Barr virus-induced lym-phomagenesis of immunodeficient humans. Int J Cancer1991, 47:510-517

23. Cannon MJ, Pisa P, Fox RI, Cooper NR: Epstein-Barr virusinduces aggressive lymphoproliferative disorders of humanB cell origin in SCID/hu chimeric mice. J Clin Invest 1990,85:1333-1337

24. Rowe M, Young LS, Crocker J, Stokes H, Henderson S,Rickinson AB: Epstein-Barr virus (EBV)-associated lym-phoproliferative disease in the SCID mouse model: Implica-tions for the pathogenesis of EBV-positive lymphomas inman. J Exp Med 1991, 173:147-158

25. Bankert RB, UmemotoT, Sugiyama Y, Chen FA, Repasky E,Yokota S: Human lung tumors, patients' peripheral bloodlymphocytes and tumor infiltrating lymphocytes propagatedin scid mice. Curr Top Microbiol Immunol 1989, 152:201-210

26. Pirruccello SJ, Bicak MS, Gordon BG, Gajl PK, Gnarra DJ,Coccia PF: Acute lymphoblastic leukemia of NK-cell lin-eage: responses to IL-2. Leuk Res 1989, 13:735-743

27. Barnes DWH, Loutit J: Spleen protection: The cellular hy-pothesis. Radiobiology Symposium. Edited by Bacq ZM.London, Butterworth, 1955, pp 134-135

28. Billingham RE, Brent L: A simple method for inducing toler-ance of skin homografts in mice. Transplant Bull 1957,4:67-71

29. Zaleski MB: Graft-versus-host reactions. Principles of Immu-nology, 2nd Edition. Edited by Rose NR, Milgron F, van OssCJ. New York, Macmillon, 1979, pp 402-413

30. Ferrara JLM, Deeg JH: Graft-versus-host disease. N EngI JMed 1991, 324:667-674

31. Gatti RA, Meuwissen HJ, Allen HD, Hong R, Good RA: Im-munological reconstitution of sex-linked lymphopenic immu-nological deficiency. Lancet 1968 fi:1366-1369

32. Nakamine H, Yokote H, Itakura T, Hayashi S, Komai N, Ta-kano Y, Saito K, Moriwaki H, Nishino E, Takenaka T, MaedaJ, Matsumori T: Non-Hodgkin's lymphoma involving thebrain. Diagnostic usefulness of stereotactic needle biopsy incombination with paraffin-section immunohistochemistry.Acta Neuropathol 1989, 78:462-471

33. Hsu SM, Raine L, Fanger H: Use of avidin-biotin-peroxidasecomplex (ABC) in immunoperoxidase techniques: A com-parison between ABC and unlabeled antibody (PAP) pro-cedures. J Histochem Cytochem 1981, 29:577-580