Anti-CD20 Antibody Therapy and Susceptibility to PneumocystisPneumonia

Waleed Elsegeiny,a,b Taylor Eddens,a,b Kong Chen,b Jay K. Kollsb

Department of Immunology, University of Pittsburgh School of Medicine (UPSOM), Pittsburgh, Pennsylvania, USAa; Richard King Mellon Foundation Institute for PediatricResearch, Children’s Hospital of Pittsburgh of UPMC, Pittsburgh, Pennsylvania, USAb

Anti-CD20 antibody therapy has been a useful medication for managing non-Hodgkin’s lymphoma as well as autoimmunediseases characterized by autoantibody generation. CD20 is expressed during most developmental stages of B lymphocytes;thus, CD20 depletion leads to B-lymphocyte deficiency. As the drug has become more widely used, there has been an in-crease in the number of case reports of patients developing Pneumocystis pneumonia. The role of anti-CD20 in Pneumocystisjirovecii infection is under debate due to the fact that most patients receiving it are on a regimen of multiple immunosuppressivemedications. To address the specific role of CD20 depletion in host immunity against Pneumocystis, we examined a murine anti-CD20 depleting antibody. We demonstrated that anti-CD20 alone is permissive for Pneumocystis infection and that anti-CD20impairs components of type II immunity, such as production of interleukin-4 (IL-4), IL-5, and IL-13 by whole-lung cells, in re-sponse to Pneumocystis murina. We also demonstrated that CD4� T cells from mice treated with anti-CD20 during Pneumocys-tis infection are incapable of mounting a protective immune response when transferred into Rag1�/� mice. Thus, CD20� cellsare critical for generating protective CD4� T-cell immune responses against this organism.

Anti-CD20 monoclonal antibodies first showed efficacy againstchemotherapy-resistant non-Hodgkin’s B-cell lymphomas in

1997 (1, 2). Since then, anti-CD20 has been used therapeuticallyagainst hematological cancers, autoimmune diseases, and post-transplant lymphoproliferative disease. CD20 is a B-lymphocyteantigen encoded by a membrane-spanning 4A family member,MS4A1. There is no known ligand for CD20; however, it is be-lieved to play a role in B-cell development and differentiation intoplasma cells and in T-cell-independent antibody (Ab) responses(3). With the increased use of anti-CD20 as a treatment, there havebeen several recent reports of patients receiving anti-CD20 andsubsequently developing infection with the opportunistic patho-gen Pneumocystis jirovecii, which may develop into a fatal pneu-monia even with antibiotic therapy (4). However, there is debateabout the precise role of anti-CD20 in conferring risk, as many ofthese patients are also on concomitant immunosuppressive drugs,thus complicating any analyses of clinical studies (5–7).

Pneumocystis is an opportunistic fungal pathogen that was origi-nally a very strong indicator that a patient had human immunodefi-ciency virus (HIV). Depletion of CD4� T cells to levels below a countof 200 per �l of blood was the primary risk factor for susceptibility toPneumocystis jirovecii pneumonia (PJP) (8, 9). The role of CD4� Tcells has been validated several times in a variety of animal models,from selective depletion of CD4� cells to the use of knockout mice(10, 11). The clearance process typically occurs either through thegeneration of effector CD4� T cells that recruit and activatephagocytes, such as macrophages, to clear the infection or by help-ing B cells to mature into Pneumocystis-specific plasma cells,which promote antibody-mediated phagocytosis (12, 13). The im-portance of B cells was first observed when it was demonstrated thatB-cell-deficient mice (�MT mice) were permissive for Pneumocystisinfection. At the time, this effect was suggested to be due to thelack of serum immunoglobulins in these mice (14). However, sub-sequent studies demonstrated that B cells play a larger role thanjust antibody generation, as Lund et al. showed that B cells wererequired for priming of CD4� T cells and for generating protective

effector and memory CD4� T cells in response to Pneumocystislung infection in mice (15). This suggested that depletion ofCD20� B cells would also lead to CD4� T-cell dysfunction andsusceptibility to Pneumocystis infection. To experimentally testthis hypothesis, we administered a murine anti-CD20 depletingantibody (5D2) to mice, followed by subsequent infection with P.murina. We found that administration of anti-CD20 conferredsusceptibility to primary Pneumocystis infection.

Furthermore, it has been reported that some patients receivinganti-CD20-containing treatment regimens for lymphoma de-velop immune reconstitution inflammatory syndrome (IRIS) af-ter receiving the last treatment (16). Thus, we next investigated theeffects of CD20 depletion on the development of IRIS in our mu-rine model. We concluded that although the pathology/lung in-jury associated with CD4� T-cell reconstitution was not influ-enced by the presence or absence of B cells, the ability of the CD4�

T cells to mount a protective immune response against Pneumo-cystis murina was in fact dependent on CD20� B cells. CD20 de-pletion did not affect the recruitment of CD4 cells to the lung, butinfected lungs had reduced type II immune responses. This studysheds some light on how anti-CD20 treatment in patients may

Received 17 December 2014 Returned for modification 31 December 2014Accepted 25 February 2015

Accepted manuscript posted online 2 March 2015

Citation Elsegeiny W, Eddens T, Chen K, Kolls JK. 2015. Anti-CD20 antibodytherapy and susceptibility to Pneumocystis pneumonia. Infect Immun83:2043–2052. doi:10.1128/IAI.03099-14.

Editor: G. S. Deepe, Jr.

Address correspondence to Jay K. Kolls, jay.kolls@chp.edu.

Supplemental material for this article may be found at http://dx.doi.org/10.1128/IAI.03099-14.

Copyright © 2015, American Society for Microbiology. All Rights Reserved.

doi:10.1128/IAI.03099-14

May 2015 Volume 83 Number 5 iai.asm.org 2043Infection and Immunity

on February 1, 2019 by guest

http://iai.asm.org/

Dow

nloaded from

affect their ability to mount a defense against Pneumocystis, and itmay be critical to monitor these patients for Pneumocystis jiroveciiinfection.

MATERIALS AND METHODSMice. Six- to 8-week-old wild-type C57BL/6J (WT), immunodeficientB6.129S7-Rag1tm1Mom/J (Rag1�/�), and B6.CB17-Prkdcscid/SzJ (SCID)mice were obtained from The Jackson Laboratory (Bar Harbor, ME).Immunodeficient B10:B6-Rag2tm1FwaIl2rgtm1Wjl (Rag2�/� Il2r��/�)mice were originally obtained from Taconic (Hudson, NY) and then bredand maintained at the University of Pittsburgh Division of LaboratoryAnimal Resources (DLAR) Facility, Children’s Hospital of Pittsburgh ofUPMC. Animals were housed in a pathogen-free environment andgiven food and water by the DLAR ad libitum. All experiments wereapproved by the University of Pittsburgh Institutional Animal Care andUse Committee.

Pneumocystis isolation, inoculum, and antigen preparation. Pneu-mocystis murina organisms were administered by oral-pharyngeal deliv-ery to Rag2�/� IL2r��/� mice, propagated for 10 to 12 weeks in vivo, andisolated from mouse lung tissue as previously described (17). Briefly,Rag2�/� IL2r��/� mice with Pneumocystis pneumonia were sacrificed,and the lungs were aseptically harvested and frozen in 1 ml of sterileDulbecco’s phosphate-buffered saline (PBS) at �80°C. To process theinoculum, frozen lungs were thawed, strained through a 70-�m filter, andpelleted by centrifugation (800 � g, 10 min, 4°C). The pellet was resus-pended in 1 ml of PBS. A 5-�l aliquot was diluted 1:10, heat fixed on aslide, and stained with Hema-3 modified Wright-Giemsa stain (FisherScientific, Pittsburgh, PA), followed by ascus counting. Pneumocystismurina asci were quantified microscopically, and the inoculum was ad-justed to 2 � 106 asci per ml. Mice were administered 100 �l (2 � 105 asci)of the inoculum by oral-pharyngeal aspiration as previously described(18). Pneumocystis protein antigen was prepared by differential centrifu-gation of the inoculum as previously described, followed by sonication of1 mg of inoculum per ml for 5 min (19).

Preparation of whole-lung cells (WLC) and antigen stimulation.Mice were infected with an inoculum of P. murina for 2 weeks. At the timeof euthanasia, mice were anesthetized by intraperitoneal injection of aketamine-xylazine cocktail and euthanized by exsanguination. Immedi-ately after, mice were perfused vascularly by 5 ml of heparinized PBSinjected into the right ventricle. The right superior and inferior lung lobeswere then harvested, minced with razor blades, and digested in 5 ml se-rum-free medium with 2 mg/ml collagenase for 90 min in a 37°C shakingincubator. The cell suspension was strained through a 70-�m filter andthen washed and resuspended in complete Dulbecco’s modified Eagle’smedium (DMEM). Red blood cells were then lysed with ammonium chlo-ride solution, washed, resuspended in 5 ml of DMEM, and counted. Atotal of 106 cells per well were plated in a 96-well round-bottom plate incomplete DMEM with 1 �g/ml P. murina antigen and 20 U/ml interleu-kin-2 (IL-2). An aliquot from each group was taken for cell analysis byflow cytometry. Plated cells were stimulated at 37°C and 5% CO2 for 72 h.Finally, supernatants were harvested for multiplex (Millipore) cytokineanalysis on a Bioplex reader (Bio-Rad).

Flow cytometric analysis. A total of 106 single cells from the mouselung were stimulated with 50 ng/ml phorbol myristate acetate (PMA)(Sigma) and 750 ng/ml ionomycin (Sigma) for 5 to 6 h. One hour after thestart of stimulation, cells were given 1 �l/ml GolgiStop (BD Pharmingen,San Diego, CA) to block cytokine secretion. Cells were surface stainedwith T-cell receptor beta (TCR�), CD4, CD8, and B220 for 15 to 30 min inPBS supplemented with 1% bovine serum albumin (BSA). Cells were thenfixed in 1% formalin and acquired for flow cytometry by an LSR-II flowcytometer (BD Biosciences, San Jose, CA), and data were analyzed usingFlowJo (Treestar).

RNA isolation and Pneumocystis quantification by RT-PCR. Theright middle lobe of the lung was harvested in 1 ml of TRIzol and homog-enized. RNA was purified and quantified as previously described (17).

Briefly, cDNA was synthesized from 1 �g whole-lung RNA via iScriptreverse transcription reagents (Bio-Rad, Hercules, CA), and real-timePCR (RT-PCR) was performed using primers and probes for the Pneu-mocystis murina large-subunit (LSU) rRNA transcript and SsoFast/SsoAdvanced probe supermix (Bio-Rad). The threshold cycle values wereconverted to copy numbers by use of a premade standard of known Pneu-mocystis LSU rRNA, as previously described (20).

Serum collection and Pneumocystis murina antigen enzyme-linkedimmunosorbent assay (ELISA). Blood was collected periodically by tailbleed and/or at the time of sacrifice by syringe from the vena cava. Coag-

FIG 1 Anti-CD20 treatment during Pneumocystis infection. (A) Flow cytom-etry of whole splenocytes from untreated mice (D0) (left) and mice at 6 daysposttreatment (D6) (right). In the upper panels, the y axis is for CD3, and thex axis is for IgM. The lower panels are gated on the IgM� CD3� quadrant of theupper panels: the y axis is for CD23, and the x axis is for CD5. Representativeimages (n � 3 per treatment) are shown. (B) C57BL/6 mice were depleted 2days prior to infection and were treated with 5D2 every 7 days. Mice weresacrificed at 2 and 4 weeks postinfection. (C) Total lung RNAs were isolated,and Pneumocystis burdens were measured by real-time PCR analysis of themitochondrial large-subunit rRNA copy number. Burdens are reported asmeans standard errors of the means (SEM) for 4 mice per group. P values inall figures are annotated as follows: *, P � 0.05; **, P � 0.01; ***, P � 0.001;****, P � 0.0001; and ns, not significant. PC, Pneumocystis.

Elsegeiny et al.

2044 iai.asm.org May 2015 Volume 83 Number 5Infection and Immunity

on February 1, 2019 by guest

http://iai.asm.org/

Dow

nloaded from

ulated blood was then centrifuged for 10 min at 10,000 � g. The serumsupernatant was collected and stored at �80°C. Maxisorb plates werecoated with 1 �g P. murina antigen in 100 �l bicarbonate coating bufferper well overnight at 4°C. Plates were blocked with 5% blotting-gradeblocker (Bio-Rad) and 1% BSA. Plates were first stained with sampleserum in a dilution series from 26 to 213 overnight at 4°C and then stainedwith murine Ig-specific horseradish peroxidase (HRP)-conjugated anti-bodies. Plates were then developed with tetramethylbenzidine (TMB)substrate for 5 to 30 min, depending on the control serum, and the reac-tion was stopped with an equal volume of 2 N H2SO4. The optical densityat 450 nm (OD450) was read using a Synergy H1 Hybrid reader (BioTek,Winooski, VT).

Antibody-mediated cell depletion. CD4� cells were depleted usingan anti-CD4 monoclonal antibody, GK1.5, as previously described (17).Mice were injected in the intraperitoneal space weekly with a 0.3-mg doseof Ab in 200 �l sterile PBS. CD4� cell depletion efficiency was assessed byflow cytometry with anti-CD4 clone RM4-5, which does not competewith GK1.5. CD20� cells were depleted using a mouse anti-mouse CD20monoclonal antibody (clone 5D2, murine IgG2a; Genentech). Mice weregiven intraperitoneal injections with 0.1-mg doses of Ab every 5 days.CD20� cell depletion efficiency was assessed by flow cytometry with anti-B220.

Preparation of splenocytes and purified CD4� and purified B220cells for adoptive transfer. Spleens harvested from C57BL/6J mice wereharvested, diced, and strained through a 70-�m filter to create a single-cell

splenocyte suspension. CD4� cells and B220 cells were purified using aStem Cell EasySep negative-selection mouse CD4� T-cell isolation kit andmouse B220 B-cell isolation kit, respectively. Cells were enumerated andresuspended in sterile PBS. Cells were resuspended at 2.5 � 106 cells/ml,and each mouse received 5 � 105 cells (200 �l) via intravenous (tail vein)injection. In testing the capacity of CD4 for clearance, cells were adop-tively transferred into mice 2 weeks prior to infection and sacrificed 4 to 6weeks after infection with Pneumocystis murina. To induce IRIS, micewere originally infected for 21 days prior to cell adoptive transfer and thensacrificed 10 days after transfer.

Pulse oximetry and lung injury qualification from BALF. The mouseblood oxygen saturation, respiratory rate, and heart rate were measuredfrom the tail by using a MouseOx pulse oximeter (Starr Life Sciences,Oakmont, PA) as previously described (18). Briefly, mice were anesthe-tized with 100 mg of intraperitoneal ketamine/kg of body weight, and thetail clip sensor was placed at the base of the tail, scanning the lateral vein.One-minute measurements were recorded, and the average values for 20stable readings were reported. The first milliliter of bronchoalveolar la-vage fluid (BALF) was harvested by repeated 0.5-ml washes with a singlefull syringe with 1 ml PBS. The BALF was centrifuged at 400 � g for 10min at 4°C to remove live cells and debris. Total protein was measuredby use of a bicinchoninic acid (BCA) protein assay kit (Pierce, Rock-ford, IL). Lactate dehydrogenase (LDH) was analyzed with an LDHactivity assay kit (BioVision, Milpitas, CA), and readouts were re-corded every 5 min for 1 h.

FIG 2 CD4 and B220 cell assessments post-anti-CD20 treatment. (A) C57BL/6 mice were administered anti-CD20 or a control and infected withPneumocystis for 2 weeks. Total lung RNAs were isolated, and Pneumocystis burdens were measured by real-time PCR analysis of the mitochondrial large-subunitrRNA copy number. (B) Whole-lung cells (WLCs) were counted from the right superior and inferior lobes. (C) The absolute lymphocyte count was determinedfrom the percent lymphocyte gate (based on forward scatter [FSC] and side scatter [SSC]) and the total lung WLC count. (D) Similarly, absolute CD4 and B220counts were calculated by determining the percent positive cells and the absolute lymphocyte count. All data are reported as means SEM for 5 mice per group.

CD20� Cells and Pneumocystis Infection

May 2015 Volume 83 Number 5 iai.asm.org 2045Infection and Immunity

on February 1, 2019 by guest

http://iai.asm.org/

Dow

nloaded from

Statistical analysis. GraphPad Prism (GraphPad Software, La Jolla,CA) one-way analysis of variance (ANOVA) with the Holm-Sidak multi-ple-comparison posttest was used to calculate P values. Nonparametricdata for three or more groups were analyzed with the Kruskal-Wallis testand the Dunn multiple-comparison posttest. Two group comparisonswere done with the multiple t test. P values are annotated as follows: *, P �0.05; **, P � 0.01; ***, P � 0.001; and ****, P � 0.0001.

RESULTSAnti-CD20 treatment induces susceptibility to Pneumocystis.First, we independently validated that 5D2 (murine anti-CD20)was capable of depleting B cells in mice. Prior to depletion (day 0),over half of the splenocytes were IgM�, and the majority were alsoCD23�. By day 6 postdepletion, this population of cells was re-duced by approximately 90% (Fig. 1A). To investigate if anti-CD20 conferred susceptibility to primary Pneumocystis infection,we administered anti-CD20 to mice, followed by Pneumocystis

murina infection (Fig. 1B). We measured P. murina burdens both2 and 4 weeks after infection, and at 2 weeks, 5D2-treated miceand isotype control mice had no differences in infectious burden.However, 4 weeks after infection, control mice began clearing in-fection, whereas 5D2-treated mice had approximately 2-log largerP. murina burdens in the lung (Fig. 1C). Thus, CD20� cells arecrucial for a protective immune response, and depletion results inhigh susceptibility to primary Pneumocystis infection. We also re-peated this experiment by assessing infection at 6 and 8 weekspostinoculation and observed that the infection burden was main-tained in anti-CD20-treated mice but was cleared in control mice(data not shown).

CD20 depletion blocks type II immune responses in the lung.To determine mechanisms by which anti-CD20 was permissive forsustained Pneumocystis infection, we assessed cellular immune re-sponses in whole-lung cells (WLCs). We harvested lungs from

FIG 3 Cytokine analysis of supernatants from Pneumocystis antigen-stimulated whole-lung cells. WLCs were cultured in the presence of P. murina antigen andIL-2 for 72 h. (A to C) All cytokines were analyzed simultaneously by a Millipore multiplex assay on a Bioplex reader. All data are reported as means SEM for5 mice per group.

Elsegeiny et al.

2046 iai.asm.org May 2015 Volume 83 Number 5Infection and Immunity

on February 1, 2019 by guest

http://iai.asm.org/

Dow

nloaded from

5D2-treated, GK1.5-treated (CD4-depleted), and isotype controlmice 2 weeks after infection. We chose the 2-week time pointbecause the infection loads were approximately equal among allthe groups at this time, so any difference in immune responsescould not be attributed to the presence of more antigen in vivo(Fig. 2A). The lungs were digested and strained to generate WLCcultures, which were analyzed by flow cytometry or stimulatedwith P. murina antigen. Total WLCs were counted so that absolutelymphocyte counts could be calculated. There was no difference inthe number of recovered WLCs; however, there was a significantreduction in the total lymphocyte count in CD20-depleted ani-mals (Fig. 2B and C). Furthermore, flow cytometry analysisshowed that there was a 98% CD4 depletion with GK1.5 admin-istration and an approximately 90% B220 depletion with 5D2(Fig. 2D).

After stimulation of WLCs with P. murina antigen for 72 h, weanalyzed effector cytokines in cell supernatants by Luminex assay.Gamma interferon (IFN-�) production by WLCs did not differ

across the groups (Fig. 3A). IL-17a was not highly produced, andalthough there was a trend toward higher levels in control mice,this was not statistically significant between the groups (Fig. 3B).However, the vehicle control group had a strong type II signature,which was substantially reduced in both the CD4- and CD20-depleted groups. Briefly, IL-4 and IL-13 were substantially re-duced in both CD4- and CD20-depleted samples. Although IL-5expression was also reduced after CD20 depletion, levels were stillsignificantly higher than those in the CD4-depleted samples (Fig.3C). However, depleting CD4� cells from CD20-depleted micefurther reduced type II responses, which were equivalent to levelswith CD4 depletion alone (data not shown). These data show thatCD20 depletion leads to a defect in effector immune responses inthe lung and that the decrease is primarily in type II immunity ofwhole-lung cells.

Anti-CD20 treatment eliminates CD4� T-cell protective re-sponses. To address whether CD20 depletion specifically altersCD4 cell effector function, we then examined whether adoptively

FIG 4 Loss of CD4� cell-specific protection after anti-CD20 treatment. (A) Mice were CD20 depleted and infected with P. murina for 2 weeks. SplenicCD4� cells were then purified and adoptively transferred into Rag1�/� mice. After homeostatic cell proliferation, mice were infected with P. murina for 4 weeks.(B) Purified splenic CD4� cells from CD20-depleted mice were stimulated with CD3/CD28 beads. Cytokines in the supernatants were analyzed simultaneouslyby Millipore multiplex assay on a Bioplex reader. (C) Total lung RNAs were isolated, and Pneumocystis burdens were measured by real-time PCR analysis of themitochondrial large-subunit rRNA copy number. Burdens are reported as means SEM for 5 mice per group.

CD20� Cells and Pneumocystis Infection

May 2015 Volume 83 Number 5 iai.asm.org 2047Infection and Immunity

on February 1, 2019 by guest

http://iai.asm.org/

Dow

nloaded from

transferring purified CD4� T cells from naive C57BL/6J WT miceinto Rag1�/� mice had any protective effects against Pneumocys-tis. Adoptive transfer of purified CD4� T cells resulted in signifi-cant reductions in P. murina lung burdens compared to those ofvehicle control mice (see Fig. S1 in the supplemental material).We next examined if CD4� T cells from 5D2-treated mice orisotype control mice could clear Pneumocystis. Briefly, 5D2-treated C57BL/6J WT mice were infected with the standard inoc-ulum of Pneumocystis murina for 2 weeks, after which CD4� cellsfrom these mice were purified and adoptively transferred toRag1�/� mice (Fig. 4A). Splenic CD4� T cells were also tested forthe capacity to respond to CD3/CD28 bead stimulation ex vivo.CD4� T cells from CD20-depleted mice had significantly lessIFN-� production and a trend toward decreased IL-5 (Fig. 4B).The cells were given 14 days to homeostatically proliferate inRag1�/� mice before infection with Pneumocystis murina. Lungswere harvested 4 weeks after infection and analyzed for P. murinaburden. Adoptive transfer of lung CD4� T cells from control miceagain resulted in control of P. murina infection, whereas lung

CD4� T cells from CD20-depleted mice were defective in clearingP. murina (Fig. 4C).

Mice with convalescent-phase immunity are resistant to an-ti-CD20-mediated susceptibility. We next determined if anti-CD20 conferred the risk of secondary infection with P. murina. Toexamine this, we generated convalescent mice by infecting themwith P. murina and allowing them 8 weeks to clear and recoverfrom the infection. After the recovery period, one group was given5D2, another group received GK1.5, and a third group received acarrier control (sterile PBS). All groups were then reinfected withthe same inoculum of P. murina, and fungal burdens were mea-sured 2 weeks after the second infection (Fig. 5A). All groups,including the CD4- and CD20-depleted groups, completely erad-icated the infection (Fig. 5B). We wanted to discern the method ofclearance, so we assessed the level of Pneumocystis murina-specificantibodies in the serum. All groups had approximately equal levelsof anti-P. murina IgG (Fig. 5C). We attributed the clearance to thepresence of anti-Pneumocystis IgG. To formally test this, we exam-ined the effector activity of these antibodies by adoptively trans-

FIG 5 Humoral immunity is sufficient for protection against Pneumocystis. (A) Wild-type C57BL/6 mice were infected with P. murina and allowed to convalescefor 6 weeks. They were then CD20 depleted and infected with P. murina for 2 weeks. (B) Total lung RNAs were isolated, and Pneumocystis burdens were measuredby real-time PCR analysis of the mitochondrial large-subunit rRNA copy number. (C) Total P. murina antigen-specific IgGs were measured by direct ELISA ofa 1:64 dilution of serum as the primary antibody and anti-mouse IgG–HRP as the secondary antibody. (D) Sera harvested from convalescent mice were adoptivelytransferred into Rag1�/� mice at the time of infection with P. murina. (E) Pneumocystis burdens following transfer of convalescent-phase serum as measured byreal-time PCR as described above. All data are reported as means SEM for 3 to 5 mice per group.

Elsegeiny et al.

2048 iai.asm.org May 2015 Volume 83 Number 5Infection and Immunity

on February 1, 2019 by guest

http://iai.asm.org/

Dow

nloaded from

ferring 200 �l of either naive or convalescent-phase serum fromWT C57BL/6 mice infected with P. murina to Rag1�/� mice,which lack T and B cells (Fig. 5D). Rag1�/� mice that receivedconvalescent-phase serum were protected from infection (Fig.5E). These studies show that anti-CD20 does not confer suscepti-bility to infection in the presence of preexisting Pneumocystis im-munity.

CD20 depletion does not influence IRIS severity. We nextdetermined if CD20� cells may contribute to IRIS. The first modelwe used to test this experiment was severe combined immunode-ficient (SCID) mice infected with Pneumocystis murina followedby adoptive transfer of whole splenocytes to induce IRIS (21). As anegative control for IRIS in this experiment, a group of SCID micereceived whole splenocytes and were depleted of CD4� T cells invivo. We then examined whether or not anti-CD20 reduced orprevented tissue pathology (Fig. 6A). We first measured bodyweight and found that mice that received anti-CD20 lost approx-imately the same percentage of total body weight as the nonde-pleted mice (Fig. 6B). However, depletion of CD4� T cells pre-vented weight loss (Fig. 6B). Total protein in the BALF, a measureof lung damage, was also significantly reduced in mice that re-

ceived anti-CD4 antibodies but increased in control and anti-CD20-treated mice (Fig. 6C). Lactate dehydrogenase activity inthe BALF, a measure of cell death in the lung, was elevated in micethat received whole splenocytes but was significantly abrogated inmice that received anti-CD4 (Fig. 6D). However, there was slightlyhigher LDH activity in mice that received anti-CD20 (Fig. 6D).These data show that IRIS requires CD4� T cells but that anti-CD20 has little effect on the development of IRIS in this model.

We next examined if CD20� cells in conjunction with CD4� Tcells affected IRIS. We also tested the hypothesis that Pneumocys-tis-experienced B cells or plasma cells would provide a protectiveeffect and mitigate IRIS. Briefly, we adoptively transferred CD4�

T cells alone, CD4� T cells plus B220 cells from naive mice, orCD4� T cells plus B220 cells from antigen-experienced (Ag Exp)mice into P. murina-infected Rag1�/� mice. Ag Exp cells werefrom mice infected with Pneumocystis murina for 14 days. Wesacrificed mice at day 10 posttransfer, as Pneumocystis burdenswere still equal between all the groups at this time point (Fig. 7A)but it was the peak day of lung injury. By day 10, we could detectanti-chitin IgM levels in the sera of the naive and antigen-experi-enced B220 cell transfer mice and anti-Pneumocystis IgG levels

FIG 6 Influence of CD20 depletion on immune reconstitution inflammatory syndrome. (A) BALB/c SCID mice were infected with P. murina for 3 weeks andthen reconstituted with WT splenocytes (SPLN), with or without CD20 or CD4 depletion for 10 days. (B) Mice were weighed before and after reconstitution todetermine the percent weight loss. (C) Total protein from BALF was measured by BCA assay. (D) Lactate dehydrogenase in the BALF was measured by a kineticassay of LDH activity. All data are reported as means SEM for 5 mice per group.

CD20� Cells and Pneumocystis Infection

May 2015 Volume 83 Number 5 iai.asm.org 2049Infection and Immunity

on February 1, 2019 by guest

http://iai.asm.org/

Dow

nloaded from

only in the Ag Exp B220 cell transfer mice (Fig. 7B and C). As inthe previous experiment, adding B220 cells, whether naive or AgExp, did not affect weight loss, blood oxygen saturation, or BALFlactate dehydrogenase (Fig. 7D to F). These data show that B cellsare dispensable for CD4� T-cell-mediated IRIS in this model.

DISCUSSION

As there is an increase in the number of incidents of opportunisticinfections involving patients receiving anti-CD20-containingtreatment regimens, there is a need to understand the effects ofanti-CD20 on the immune response to infection. We pursued thisquestion by developing an in vivo mouse model of CD20 depletionin an attempt to recapitulate the clinical scenario of anti-CD20

treatment. We were first able to observe that anti-CD20 alone wascapable of inducing susceptibility to Pneumocystis infection bydiminishing the immune response to Pneumocystis in the lung.WLCs showed reduced type II responses, and CD4� T cells spe-cifically were not capable of providing protective responses in anadoptive transfer model. These data suggest that anti-CD20 inhib-its T-cell priming against P. murina. (22). It is still not known whyCD20� B cells specifically are required to prime CD4� cell re-sponses to Pneumocystis. Previous studies have shown that bothCD40 and major histocompatibility complex (MHC) class II mol-ecules on B cells are critical for CD4� T-cell priming; however,why other antigen-presenting cells are not capable is still a ques-tion that needs to be addressed (13). Interestingly, the absence of

FIG 7 Influence of naive and antigen-experienced B220 cells on immune reconstitution inflammatory syndrome. (A) Total lung RNAs from mice reconstitutedwith a combination of CD4� T cells and naive or antigen-experienced B220 cells were isolated, and Pneumocystis burdens were measured by real-time PCRanalysis of the mitochondrial large-subunit rRNA copy number. (B and C) Anti-chitin IgM (B) and anti-P. murina IgG (C) in serum were assessed by directELISA with HRP-conjugated anti-IgM and anti-IgG secondary antibodies, respectively. (D) Mice were weighed before and after reconstitution to determine thepercent weight loss. (E) Mouse blood oxygenation was measured immediately prior to sacrifice by use of a pulse oximeter. (F) Lactate dehydrogenase in the BALFwas measured by a kinetic assay of LDH activity. All data are reported as means SEM for 4 or 5 mice per group.

Elsegeiny et al.

2050 iai.asm.org May 2015 Volume 83 Number 5Infection and Immunity

on February 1, 2019 by guest

http://iai.asm.org/

Dow

nloaded from

CD20� cells during Pneumocystis infection may lead to clonal de-letion of Pneumocystis-specific CD4� cells, since they are not ascapable of clearing Pneumocystis as CD4� cells from naive mice.This may be due to a lack of CD4� T-cell costimulation by B cellsduring Pneumocystis infection.

While the previous experiments were conducted with naivemice, anti-CD20-treated patients are likely not naive to Pneumo-cystis jirovecii, as most individuals are seropositive for anti-Pneu-mocystis antibodies (19, 23). The potential reservoirs for Pneumo-cystis are still unknown, but some research indicates the possibilitythat many diseased and healthy patients may be colonized withPneumocystis. Some studies also suggest that hosts may be tran-siently exposed or asymptomatically carry Pneumocystis through-out their lives (24). We observed that convalescent mice werecapable of clearing Pneumocystis without the aid of CD4� orCD20� cells, most likely through antibody-mediated protection.Since memory CD4� cells and antibodies are enough to clearPneumocystis, these data raise the question of why immunosup-pressed patients acquire Pneumocystis jirovecii infections. Anti-CD20 antibody therapy in patients alone has not been shown todecrease immunoglobulin levels; however, if this therapy iscombined with other immunosuppressive therapies, hypogam-maglobulinemia can occur (25). It has also been demonstratedthat Pneumocystis spp. have a dynamic extracellular proteome as aresult of varying their major surface glycoproteins, which may beused to evade the host immune response (26, 27). Reports of largegenetic variability between Pneumocystis jirovecii isolates also sug-gest the possibility that patients are simply being exposed to newstrains of Pneumocystis (22, 28).

IRIS is a condition that was first observed with the advent ofhighly active antiretroviral therapy (HAART), which dramaticallysuppresses HIV replication (29). This suppression leads to an ex-pedient reconstitution of CD4� T cells, which then react to anaccumulation of viable infections and/or residual microbial anti-gens acquired during immunosuppression. As a result, the hostcan experience a massive amount of inflammation in multipleorgans, primarily the lungs and central nervous system (CNS) (30,31). This immunopathology is mediated through CD4� T cells(32), but it was previously shown that other cell types, such as CD8cells, modulate the pathology by increasing the Treg:Teffector ratio(33). Previous studies have shown that natural antibodies mayreduce pathology (18); however, the direct role of B cells duringIRIS has yet to be studied. Thus, as B cells are required for T-celleffector function in protection, we hypothesized that CD20� cellsmay also play a role in the inflammatory pathology of T-cell-mediated IRIS. We observed that the absence or presence ofCD20� cells did not exacerbate the pathology in the model studiedhere. We hypothesize that this is due to nonspecific CD4� T-cellactivation through homeostatic proliferation. Thus, we concludethat CD20� cells are dispensable for CD4-mediated IRIS.

Thus, in conclusion, anti-CD20 confers a risk of primary in-fection with P. murina, but we did not observe a risk of secondaryinfection in mice that had preexisting humoral immunity. In ad-dition, mice treated with anti-CD20 had reduced whole-lung-celltype II responses to Pneumocystis, and CD4 cells from depletedmice had an intrinsic impairment in the ability to clear Pneumo-cystis in our adoptive transfer model. These data suggest that clin-ical Pneumocystis jirovecii pneumonia may be due to concomitantimmunosuppression or to patients acquiring antigenically dis-tinct strains of Pneumocystis that preexisting humoral immunity is

ineffective at preventing. Lastly, anti-CD20 did not affect the se-verity of CD4 T-cell-mediated IRIS in this model.

ACKNOWLEDGMENTS

We acknowledge support from the following NIH grants: R01HL062052and R01HL061271 (to J.K.K.). This research was partially supported by acampus-supported fellowship under the Training Program in Autoim-munity (grant 5T32AI089443 to P. Morel; awarded to W.E.) at the Uni-versity of Pittsburgh School of Medicine.

We thank Andrew Chan for providing 5D2 for these studies.

REFERENCES1. Reff ME, Carner K, Chambers KS, Chinn PC, Leonard JE, Raab R,

Newman RA, Hanna N, Anderson DR. 1994. Depletion of B-cells in vivoby a chimeric mouse human monoclonal antibody to CD20. Blood 83:435– 445.

2. Scott SD. 1998. Rituximab: a new therapeutic monoclonal antibody fornon-Hodgkin’s lymphoma. Cancer Pract 6:195–197. http://dx.doi.org/10.1046/j.1523-5394.1998.006003195.x.

3. Kuijpers TW, Bende RJ, Baars PA, Grummels A, Derks IA, DolmanKM, Beaumont T, Tedder TF, van Noesel CJ, Eldering E, van Lier RA.2010. CD20 deficiency in humans results in impaired T cell-independentantibody responses. J Clin Invest 120:214 –222. http://dx.doi.org/10.1172/JCI40231.

4. Deborska-Materkowska D, Kozinska-Przybyl O, Mikaszewska-Sokolewicz M, Durlik M. 2014. Fatal late-onset Pneumocystis pneumo-nia after rituximab: administration for posttransplantation recurrence offocal segmental glomerulosclerosis— case report. Transplant Proc 46:2908 –2911. http://dx.doi.org/10.1016/j.transproceed.2014.09.010.

5. Farkas JD, Clouser RD, Garrison GW. 2014. Pneumocystis pneumoniafollowing rituximab. Chest 145:663– 664. http://dx.doi.org/10.1378/chest.13-2539.

6. Martin-Garrido I, Carmona EM, Specks U, Limper AH. 2013. Pneumo-cystis pneumonia in patients treated with rituximab. Chest 144:258 –265.http://dx.doi.org/10.1378/chest.12-0477.

7. Kurokawa T, Kaya H, Yoshida T. 2010. Two cases of Pneumocystisjirovecii pneumonia with non-Hodgkin’s lymphoma after CHOP-basedchemotherapy containing rituximab. J Clin Exp Hematopathol 50:159 –162. http://dx.doi.org/10.3960/jslrt.50.159.

8. Phair J, Munoz A, Detels R, Kaslow R, Rinaldo C, Saah A. 1990. The riskof Pneumocystis carinii pneumonia among men infected with humanimmunodeficiency virus type 1. Multicenter AIDS Cohort Study Group.N Engl J Med 322:161–165.

9. Stansell JD, Osmond DH, Charlebois E, LaVange L, Wallace JM,Alexander BV, Glassroth J, Kvale PA, Rosen MJ, Reichman LB, TurnerJR, Hopewell PC. 1997. Predictors of Pneumocystis carinii pneumonia inHIV-infected persons. Pulmonary Complications of HIV Infection StudyGroup. Am J Respir Crit Care Med 155:60 – 66.

10. Shellito J, Suzara VV, Blumenfeld W, Beck JM, Steger HJ, Ermak TH.1990. A new model of Pneumocystis carinii infection in mice selectivelydepleted of helper T lymphocytes. J Clin Invest 85:1686 –1693. http://dx.doi.org/10.1172/JCI114621.

11. Hanano R, Reifenberg K, Kaufmann SH. 1996. Naturally acquired Pneu-mocystis carinii pneumonia in gene disruption mutant mice: roles of dis-tinct T-cell populations in infection. Infect Immun 64:3201–3209.

12. Garvy BA, Wiley JA, Gigliotti F, Harmsen AG. 1997. Protection againstPneumocystis carinii pneumonia by antibodies generated from either Thelper 1 or T helper 2 responses. Infect Immun 65:5052–5056.

13. Lund FE, Schuer K, Hollifield M, Randall TD, Garvy BA. 2003. Clear-ance of Pneumocystis carinii in mice is dependent on B-cells but not on P.carinii-specific antibody. J Immunol 171:1423–1430. http://dx.doi.org/10.4049/jimmunol.171.3.1423.

14. Marcotte H, Levesque D, Delanay K, Bourgeault A, de la Durantaye R,Brochu S, Lavoie MC. 1996. Pneumocystis carinii infection in transgenicB-cell-deficient mice. J Infect Dis 173:1034 –1037. http://dx.doi.org/10.1093/infdis/173.4.1034.

15. Lund FE, Hollifield M, Schuer K, Lines JL, Randall TD, Garvy BA. 2006.B-cells are required for generation of protective effector and memoryCD4� cells in response to Pneumocystis lung infection. J Immunol 176:6147– 6154. http://dx.doi.org/10.4049/jimmunol.176.10.6147.

16. Canaani J, Amit S, Ben-Ezra J, Cohen Pour M, Sarid N, Lerman H,

CD20� Cells and Pneumocystis Infection

May 2015 Volume 83 Number 5 iai.asm.org 2051Infection and Immunity

on February 1, 2019 by guest

http://iai.asm.org/

Dow

nloaded from

Perry C, Polliack A, Naparstek E, Herishanu Y. 2013. Paradoxicalimmune reconstitution inflammatory syndrome associated with ritux-imab-containing regimen in a patient with lymphoma. J Clin Oncol 31:e178 – e180. http://dx.doi.org/10.1200/JCO.2012.43.6188.

17. Zheng M, Shellito JE, Marrero L, Zhong Q, Julian S, Ye P, Wallace V,Schwarzenberger P, Kolls JK. 2001. CD4� T cell-independent vaccina-tion against Pneumocystis carinii in mice. J Clin Invest 108:1469 –1474.http://dx.doi.org/10.1172/JCI13826.

18. Ricks DM, Chen K, Zheng M, Steele C, Kolls JK. 2013. Dectin immu-noadhesins and Pneumocystis pneumonia. Infect Immun 81:3451–3462.http://dx.doi.org/10.1128/IAI.00136-13.

19. Shellito JE, Tate C, Ruan S, Rolls J. 2000. Murine CD4� T lymphocytesubsets and host defense against Pneumocystis carinii. J Infect Dis 181:2011–2017. http://dx.doi.org/10.1086/315487.

20. Steele C, Marrero L, Swain S, Harmsen AG, Zheng M, Brown GD,Gordon S, Shellito JE, Kolls JK. 2003. Alveolar macrophage-mediatedkilling of Pneumocystis carinii f. sp. muris involves molecular recognitionby the Dectin-1 beta-glucan receptor. J Exp Med 198:1677–1688. http://dx.doi.org/10.1084/jem.20030932.

21. McKinley L, Logar AJ, McAllister F, Zheng M, Steele C, Kolls JK. 2006.Regulatory T cells dampen pulmonary inflammation and lung injury in ananimal model of Pneumocystis pneumonia. J Immunol 177:6215– 6226.http://dx.doi.org/10.4049/jimmunol.177.9.6215.

22. Rostved AA, Sassi M, Kurtzhals JA, Sorensen SS, Rasmussen A, Ross C,Gogineni E, Huber C, Kutty G, Kovacs JA, Helweg-Larsen J. 2013. Out-break of Pneumocystis pneumonia in renal and liver transplant patientscaused by genotypically distinct strains of Pneumocystis jirovecii. Transplan-tation 96:834–842. http://dx.doi.org/10.1097/TP.0b013e3182a1618c.

23. Ponce CA, Gallo M, Bustamante R, Vargas SL. 2010. Pneumocystiscolonization is highly prevalent in the autopsied lungs of the general pop-ulation. Clin Infect Dis 50:347–353. http://dx.doi.org/10.1086/649868.

24. Shteinberg M, Shaked-Mishan P, Kinarti A, Abramovitch A, Amital A,Jacobi A, Kolup Feldmann AE, Shiner M, Gershtein V, Weber G, AdirY. 2014. Asymptomatic carriage of Pneumocystis jirovecii and cytomeg-alovirus in lungs of immunocompetent patients. Lung 192:875– 879. http://dx.doi.org/10.1007/s00408-014-9644-z.

25. Marco H, Smith R, Jones R, Guerry M-J, Catapano F, Burns S,Chaudhry A, Smith K, Jayne D. 2014. The effect of rituximab therapy onimmunoglobulin levels in patients with multisystem autoimmune disease.BMC Musculoskelet Disord 15:178. http://dx.doi.org/10.1186/1471-2474-15-178.

26. Gigliotti F. 1992. Host species-specific antigenic variation of a mannosy-lated surface glycoprotein of Pneumocystis carinii. J Infect Dis 165:329 –336. http://dx.doi.org/10.1093/infdis/165.2.329.

27. Sunkin SM, Stringer JR. 1996. Translocation of surface antigen genes toa unique telomeric expression site in Pneumocystis carinii. Mol Microbiol19:283–295. http://dx.doi.org/10.1046/j.1365-2958.1996.375905.x.

28. Sun L, Huang M, Wang J, Xue F, Hong C, Guo Z, Gu J. 2015.Genotyping of Pneumocystis jirovecii isolates from human immunodefi-ciency virus-negative patients in China. Infect Genet Evol 31C:209 –215.http://dx.doi.org/10.1016/j.meegid.2015.01.021.

29. Shelburne SA, Visnegarwala F, Darcourt J, Graviss EA, Giordano TP,White AC, Jr, Hamill RJ. 2005. Incidence and risk factors for immunereconstitution inflammatory syndrome during highly active antiretroviraltherapy. AIDS (London, England) 19:399 – 406. http://dx.doi.org/10.1097/01.aids.0000161769.06158.8a.

30. Lai RP, Nakiwala JK, Meintjes G, Wilkinson RJ. 2013. The immuno-pathogenesis of the HIV tuberculosis immune reconstitution inflamma-tory syndrome. Eur J Immunol 43:1995–2002. http://dx.doi.org/10.1002/eji.201343632.

31. Longley N, Harrison TS, Jarvis JN. 2013. Cryptococcal immune recon-stitution inflammatory syndrome. Curr Opin Infect Dis 26:26 –34. http://dx.doi.org/10.1097/QCO.0b013e32835c21d1.

32. Antonelli LR, Mahnke Y, Hodge JN, Porter BO, Barber DL, DerSimo-nian R, Greenwald JH, Roby G, Mican J, Sher A, Roederer M, Sereti I.2010. Elevated frequencies of highly activated CD4� T cells in HIV�patients developing immune reconstitution inflammatory syndrome.Blood 116:3818 –3827. http://dx.doi.org/10.1182/blood-2010-05-285080.

33. Swain SD, Meissner NN, Harmsen AG. 2006. CD8 T cells modulate CD4T-cell and eosinophil-mediated pulmonary pathology in Pneumocystispneumonia in B-cell-deficient mice. Am J Pathol 168:466 – 475. http://dx.doi.org/10.2353/ajpath.2006.050724.

Elsegeiny et al.

2052 iai.asm.org May 2015 Volume 83 Number 5Infection and Immunity

on February 1, 2019 by guest

http://iai.asm.org/

Dow

nloaded from