Sensitive and reproducible quantitation of mucosal HIV-1 ... · Journal of Virological Methods 95 (2001) 65–79 Sensitive and reproducible quantitation of mucosal HIV-1 RNA and DNA

Journal of Virological Methods 95 (2001) 65–79

Sensitive and reproducible quantitation of mucosal HIV-1RNA and DNA viral burden in patients with detectable and

undetectable plasma viral HIV-1 RNA using endoscopicbiopsies

Pete A. Anton a,b,*, Michael A. Poles a,b, Julie Elliott a,b, S.H. Mao b,Ian McGowan a,b,1, Heinz-Josef Lenz c, Irvin S.Y. Chen b

a Di�ision of Digesti�e Diseases, Department of Medicine, MRL 2734, UCLA School of Medicine,675 Charles E. Young Dri�e South, Los Angeles, CA 90095, USA

b UCLA AIDS Institute, Los Angeles, CA, USAc Department of Hematology/Oncology, USC School of Medicine, Los Angeles, CA, USA

Received 2 November 2000; received in revised form 22 February 2001; accepted 23 February 2001

Abstract

Mucosal tissue is the main portal of entry for HIV-1 infection and, in macaques, has been demonstrated to be asignificant compartment for viral replication and CD4+ T lymphocyte depletion. Quantitating tissue viral burden inaddition to plasma viral load provides insights into HIV-1 pathogenesis and an additional means to gaugeantiretroviral response. The aim of this study was to develop reliable, reproducible, and sensitive assays to quantitatetissue viral burden of HIV-1 RNA and DNA using 1–3 endoscopically acquired, rectosigmoid biopsies. Total DNAand RNA were simultaneously extracted following homogenization from the same tissue samples. Quantitativepolymerase chain reaction (PCR) assay in the HIV-1 LTR region was used to detect viral DNA and RT-PCR forviral RNA. It was determined that HIV-1 RNA and DNA can be reproducibly quantified from a single rectosigmoidbiopsy with minimal intra-assay or intra-patient variability. These results reflect high recovery of extracted nucleicacids with calculated results accurately reflecting in vivo levels. The techniques outlined differ from currently availableapproaches by incorporating control standards to identify loss or degradation of RNA and DNA from acquisitionthrough the in vitro assay and permit extraction with high yields of RNA and DNA from the same tissue sample.© 2001 Elsevier Science B.V. All rights reserved.

Keywords: Quantitative; HIV-1; Mucosal; Viral load; Tissue; Gastrointestinal; Gastrointestinal associated lymphoid tissue

www.elsevier.com/locate/jviromet

1. Introduction

The intestinal mucosa contains most of thebody’s lymphoid tissue (Cerf-Bensussan and Guy-

* Corresponding author. Tel.: +1-310-2065797; fax: +1-310-2068824.

E-mail address: [email protected] (P.A. Anton).1 Present address: Gilead Sciences, Foster City, CA, USA.

0166-0934/01/$ - see front matter © 2001 Elsevier Science B.V. All rights reserved.

PII: S 0166 -0934 (01 )00295 -6

P.A. Anton et al. / Journal of Virological Methods 95 (2001) 65–7966

Grand, 1991; Mowat and Viney, 1997) andtherefore constitutes an immense compartmentfor persistent HIV-1 infection. Assays to quan-tify mucosal tissue viral burden have not beenfully characterized for reproducibility nor havethey been validated (Cavert, 1998). In contrast,measurements of plasma viremia and, to a lesserextent, peripheral blood mononuclear cells(PBMCs) or isolated lymphoid cell populationshave been studied more rigorously (Cohen et al.,1997; Hockett et al., 1999).

Mucosal CD4+ T lymphocytes are predomi-nately of the activated, memory phenotype(Zeitz et al., 1988; Ullrich et al., 1990) and arehighly vulnerable to infection (Lapenta et al.,1999; Anton et al., 2000). Other cells present inthe tissue sample but not routinely recoveredduring isolation procedures either carry or arealso infected with HIV-1 (Fantini et al., 1993;Geijtenbeek et al., 2000). Successful treatment ofHIV-1 will eventually require clearance of thevirus not only from the blood, but also from thegut and other lymphoid tissue compartments.

Although there is now an increasing frequencyof reports characterizing virologic parameters intissue compartments (mucosal and lymph node)(Pantaleo et al., 1991; Sei et al., 1994; Haase etal., 1996; Cavert et al., 1997; Andersson et al.,1998; Fackler et al., 1998; Kotler et al., 1998;Lafeuillade et al., 1998; Notermans et al., 1998;Perrin et al., 1998; Hockett et al., 1999; Dybulet al., 2000), there has been little methodologicalvalidation of the reproducibility of these assays.Sensitive and reproducible virologic assessmentof mucosal tissue would be a useful means ofassessing the contributions of tissue compart-ments to ongoing viral replication. This type ofassay would be of use in longitudinal monitoringof changes following therapy, such as in clinicaltrials. As evidence emerges that resistance pat-terns and virologic profiles may differ in tissueand blood, the interest in assessing these com-partments will likely increase (al-Mulla et al.,1997; Cohen et al., 1997; Poles et al., 2001).

Recent studies have convincingly demon-strated that viral replication persists in lymphoidtissue in patients with prolonged periods of un-

detectable levels of HIV-1 in the plasma com-partment (Pantaleo et al., 1991; Haase et al.,1996). Given that only 2% of the body’slymphocytes are present in blood at any onetime, investigations of the large tissue reservoirsof these infectable targets have drawn increasingattention. Most reports have focused on lymphnodes and tonsils and have documented not onlyongoing infection but also significant numbers ofcells infected latently (Haase et al., 1996; Wonget al., 1997).

The gastrointestinal associated lymphoid tissue(GALT) is the body’s largest and most extensivelymphoid organ and is the site of extensivespread of the immunodeficiency virus and CD4+

T cell depletion, regardless of the route of infec-tion (Veazey et al., 1998; Kewenig et al., 1999).Several groups have demonstrated the presenceof HIV-1 in human gut tissue and the vulnera-bility of mucosal T cells to infection (Fantini etal., 1993; Kotler et al., 1998; Lapenta et al.,1999; Anton et al., 2000). Many mucosal celltypes such as epithelial and dendritic cells can beinfected potentially by HIV-1 or carry the virusand need to be included with mucosal CD4+ Tlymphocytes in estimates of tissue viral burden(Cavert, 1998). Mucosal biopsies obtained dur-ing endoscopy are extremely safe, painless and aminimally invasive means of obtaining mucosaltissue on a repeated basis, for quantitation. Al-though other groups have reported measures ofHIV-1 viral activity in this compartment, no val-idated methods have been described that meetthe quality controls issue described above andalso account for potential regional variation inlymphoid populations.

Using endoscopic biopsies obtained using flex-ible sigmoidoscopy, a reproducible and sensitivemethod to quantify copies of HIV-1 RNA andDNA from the same endoscopic biopsy has beendeveloped. These assays are able to provide highrecovery of nucleic acids in a manner that accu-rately reflects tissue viral burden in vivo. Theresults from one biopsy are representative ofother samples obtained concurrently at the samelevel with low intra-patient/inter-biopsy variabil-ity and minimal intra-assay variability.

P.A. Anton et al. / Journal of Virological Methods 95 (2001) 65–79 67

2. Materials and methods

2.1. Study subjects

Intestinal tissue was collected from 26 HIV-pos-itive volunteers and 5 healthy HIV-negative indi-viduals undergoing elective endoscopy for ahistory of blood in stool or for routine polypscreening. Subject characteristics including gen-der, age, CDC class, concurrent CD4+ T cellcount, concurrent plasma viral load, medications,and years of HIV-1 infection were collected attime of endoscopy and are shown in Table 1.Informed consent was obtained from all patientsand the study was approved by the UCLA Hu-man Subjects Protection Committee.

2.2. Endoscopy and biopsy acquisition

For endoscopic biopsies, a standardized site inthe rectosigmoid colon, 30 cm from the analmargin, was used for all sampling to avoid poten-tial regional variation. Biopsies were collected us-ing large cup endoscopic biopsy forceps(Microvasive Radial Jaw c1589, outside diame-ter 3.3 mm, Boston, MA) and immediately placedin liquid nitrogen. Tissue was also collected at thesame level for histopathological assessment (usingroutine hematoxylin and eosin staining) to ex-clude confounding conditions such as inflamma-tion secondary to infectious or traumatic proctitis.All biopsies were reviewed in a blinded fashion bya gastrointestinal surgical pathologist. Phle-

Table 1Clinical characteristics of HIV+ study subjects

CD4 Antiretroviral medicationbPlasma viralYrs HIV+SexPatient ID no.loada

34 900345 12M 506 Amprenavir, hydroxyurea, stavudine8740355 Indinavir sulfate, lamivudine, ritonavir, stavudineM 12 2725338365 Zidovudine+ lamivudineM 11 629

Abacavir, nevirapine3661228366 13M11 177 366 276 Amprenavir, zidovudine+ lamivudine, ritonavirM387

407 Hydroxyurea, efavirenz, didanosine, stavudine11 45820914M60 000300 NKNKcM34360 000281 NKM 4 230

F 8 180 213 413 Ritonavir, indinavir sulfate, lamivudine, stavudine, efavirenz,279adefovir dipivoxil

M 8104 824 �50 Lamivudine, stavudine, indinavir sulfate162 Saquinavir ritonavir, nevirapine, didanosine�5016012M

39411 �50M142 Indinavir sulfate, zidovudine+ lamivudine109 287M 4 Zidovudine+ lamivudine, saquinavir, nelfinavir�50

11M101 Lamivudine, ritonavir, stavudine, saquinavir�5042410 240 13 836 Stavudine, lamivudine, indinavir sulfateM716I

M 43 8007 Saquinavir, didanosine, nelfinavir360 663M 8371 439 3438 Zidovudine+ lamivudine, nelfinavirM 12376 137 189 555 Lamivudine, saquinavir, ritonavir, nevirapine, stavudineM 10375 605 41 195 Lamivudine, nelfinavir, stavudine

Stavudine, nelfinavir, ritonavir, efavirenz42 164241389 10MM 6396 154 Lamivudine, saquinavir23 484

12M399 Abacavir, hydroxyurea, didanosine, nevirapine14 225467M 7405 414 15 392 Zidovudine+ lamivudine, nevirapine

1261 Indinavir sulfate, lamivudine, ritonavir, stavudine406 M 4 478248M 59 289 Lamivudine, saquinavir, stavudine16416204M 12 514 Zidovudine+ lamivudine, indinavir sulfate15361

a Viral load at time of endoscopy.b Antiretroviral therapy at time of endoscopy.c Not known.


botomy was undertaken immediately before en-doscopy; blood was collected in EDTA-contain-ing tubes.

2.3. Protocol for the isolation of RNA and DNAfrom endoscopic biopsies

Individual endoscopic biopsies (average 10–12�g each) were collected into 2 ml pre-labeledcryo-vials (Nalgene, Rochester, NY) and immedi-ately snap frozen in liquid nitrogen. Biopsies arekept frozen at −80°C until nucleic acid extrac-tion. Total RNA and DNA are simultaneouslyextracted from each individual biopsy using amodification of the Trizol isolation protocol pre-viously reported (Mitsuyasu et al., 2000). Theaddition and tracking of exogenous RNA andDNA controls using HIV-1 LTR RNA (forseronegative samples only) or modified cy-clophilin RNA (for seropositive samples) andfirefly luciferase DNA, allow for the determina-tion of nucleic acid recovery throughout the ex-traction procedure.

Briefly, 140 �l of urea lysis buffer (ULB) and 70�l of Rnase/Dnase free water (Geno-Technolo-gies, St Louis, MO) containing 108 copies ofmodified cyclophilin ‘armored RNA’ (Ambion,Austin, TX) for seropositive samples or gradedamounts of HIV-1 LTR RNA for seronegativesamples and 13 300 copies of firefly luciferaseplasmid-derived DNA copies are added to thefrozen biopsy in a sterile 50 ml conical centrifuge.The tissue is immediately homogenized using aPowergen 125 tissue homogenizer (Fisher Scien-tific, Pittsburg, PA) for �1 min before the addi-tion of 500 �l of saturated phenol (Fisher Biotech,Pittsburg, PA). The homogenate is subjected to arapid spin (1 min at 2000 rpm) to reduce foamingand 355 �l is separated to a 2 ml Eppendorf tube.Total RNA is isolated from this fraction by theaddition of 1 ml of Trizol reagent (Gibco LifeTechnologies, Rockville, MD). Following a 5 minincubation at room temperature, 200 �l of chloro-form is added and the homogenate is centrifugedat 4°C for 15 min. The aqueous (RNA contain-ing) phase is separated and diluted with an equalvolume of 70% ethanol before applying to anRNeasy column (Qiagen, Valencia, CA) for fur-

ther extraction, as per the manufacturer’s proto-col. In order to ensure that no DNAcontamination is present, the RNA is treated for40 min at 37°C with Dnase (Promega, Madison,WI) followed by re-extraction with phenol–chlo-roform and ethanol precipitation. The RNA isre-dissolved in 30–50 �l of DEPC water and theyield determined spectophotometrically. An aver-age biopsy is found to yield �10–20 �g of totalRNA (data not shown). Agarose gel electrophore-sis is used to confirm that the RNA contains both28s and 18s species of ribosomal RNA with littleor no evidence of degradation. In addition, sam-ples of non-reverse transcribed RNA are sub-jected to PCR amplification with DNA Taqpolymerase to demonstrate that there is no evi-dence of DNA contamination.

DNA is isolated from the remaining ULB ho-mogenate by a series of phenol–chloroform (×3)and chloroform (×2) extractions followed byprecipitation at −20°C in absolute ethanol(2.5× volumes) with 5 M NaCl (0.1× volume)for at least 30 min. The precipitate is washed oncewith 70% ethanol, air-dried and dissolved in 50 �lof Rnase/Dnase free water. The yield is deter-mined spectrophotometrically and an averagebiopsy is found to yield between 20 and 40 �g ofDNA (data not shown).

2.4. Tissue HIV-1 RNA and DNAamplification/quantitation

2.4.1. HIV-1 RNA amplificationHIV-1 RNA is quantified by RT-PCR using the

Thermo-stable RTth reverse transcriptase RNAPCR kit (PE Biosystems, Foster City, CA). Gene-specific reverse transcription of 100 ng of totalRNA is accomplished using the HIV-1 LTRprimer AA55 (100 ng/�l) (5�-CTGCTAGA-GATTTTCCACACTGAC-3�) and PCR is carriedout with P-32 end-labeled HIV-1 LTR primer 667(5�-GGCTAACTAGGGAACCCACTG-3�). RNAis then subjected to 30 cycles of amplification in a4800 series Perkin Elmer thermal cycler with eachcycle consisting of a 1 min denaturation step at94°C followed by a 2 min annealing step at 65°C.RNA standards (10–3000 copies) constructed bythe in vitro transcription of cloned HIV-1 DNA


are included with each series of reactions. Stan-dards are diluted in 100 ng of HIV-1-seronega-tive tissue RNA prepared by our extractionprotocol. Radiolabeled PCR products are re-solved on a 6% polyacrylamide gel followed byautoradiography on phosphorscreens (MolecularDynamics, Sunnyvale, CA). Screens are scannedwith a 445SI Phosphoimager (Molecular Dynam-ics, Sunnyvale, CA) and quantified with the useof IMAGEQUANT software. Pixel volumes areextrapolated to copy number based on the com-parative values obtained from the linear portionof the generated standard curve. Copies are ex-pressed per �g of tissue total RNA; sensitivitywas 10 copies per PCR reaction.

To determine whether total RNA extractedfrom tissue biopsies could be added to standard-ized kits used to quantitate plasma viral burden,tissue RNA was assayed in the Amplicor®

plasma kit. Total RNA extracted from mucosalbiopsies of HIV-1-seropositive subjects anddemonstrated to have quantifiable levels of HIV-1 LTR RNA in tissue was then serially diluted inHIV-1-seronegative plasma and quantified usingthe Amplicor® plasma kit. This includes a secondextraction step which is incorporated into theAmplicor® kit.

2.4.2. HIV-1 DNA amplificationHIV-1 DNA is quantified by PCR using the

HIV-1 LTR primers AA55 and 667 describedabove. In general, 100 ng of tissue DNA, in thepresence of 1.5 units of DNA Taq Polymeraseand 0.5 mM MgCl2, is subjected to 30 cycles ofPCR under the conditions previously described.DNA standards (10–3000 copies) constructed ofcloned HIV-1 DNA (Pykjrcsf) diluted in 100 ngof seronegative tissue DNA are run simulta-neously. 10–20 ng of the tissue DNA sample isamplified using the �-globin primers LA1 (5�-ACACAACTGTGTTCACTAGC-3�) and LA2(5�-CAACTTCATCCACGTTCACC-3�) and theresults extrapolated against a standard curveconstructed from genomic DNA quantified spec-trophotometrically. HIV-1 DNA is normalized to2×106 copies of �-globin to permit comparisonto be made between biopsies; sensitivity was 10copies per PCR reaction.

2.5. Mucosal co-culture studies

Mucosal biopsies obtained by endoscopy fromHIV-1 infected patients were collected into 10–15 ml of RPMI supplemented with antibiotic/an-timycotic solution (GIBCO BRL, Rockville,MD). The mucosal biopsies were rinsed threetimes in 10–15 ml of sterile PBS, transferred to a60×15 mm2 petri dish and teased apart usingtwo 18-gauge needles. The minced tissue wasmaintained for 3 days in a humidified tissue cul-ture incubator at 37°C with 5% CO2 in RPMIsupplemented with 10 IU per ml of IL-2 (Am-gen, Thousand Oaks, CA) and 20% fetal calfserum (Gemini Bioproducts, Calabasas, CA)with 100 �g/ml each of penicillin and strepto-mycin (Omega Scientific, Tarzana, CA).

After 3 days in culture, the minced biopsypieces and isolated mucosal mononuclear cellswere re-plated in 10 ml RPMI supplementedwith 10 IU of IL-2 and 20% fetal calf serum)with 100 �g/ml each of penicillin and strepto-mycin in the presence of 5×106 3-day PHA-stimulated PBMCs. Supernatant from thesecultures was analyzed for HIV-1 p24 protein byELISA (Coulter Inc, Miami, FL). Results areexpressed in terms of ng of p24 produced per mlfrom four biopsies.

2.6. Statistical methodology

Random effects models were used to assess thecomponents of variability on the log10 scale. Re-sults are reported as a maximum likelihood esti-mate and a 95% confidence interval for thestandard deviation of each random event compo-nent. All analyses were performed using SAS sys-tem software (SAS Institute, Cary, NC).

3. Results

3.1. Characterization of mucosal tissue samples

The endoscopic appearance of the HIV-1-seronegative subject’s mucosa was unremarkable


Fig. 1. Representative colonic histology and biopsy. Histologic section from normal-appearing colon biopsy and immunohistochem-ical staining of CD4+ lymphocytes. (a) shows hematoxylin and eosin staining (40× ) of normal colon structure with parallelarrangement of colonic glands with intact epithelia and scattered intraepithelial lymphocytes overlying lamina propria withsignificant numbers of normally resident lymphocytes. Lumen is at the top. (b) demonstrates a representative luminal edge of onecolonic crypt stained with anti-CD4 antibody counterstained with peroxidase to indicate the marked numbers of CD4+ lymphocytespresent in the normal colon. (c) is included to demonstrate the size of tissue sample used for RNA and DNA extraction andamplification.

and the biopsies were without pathologic diagno-sis. The endoscopic appearance of the HIV-1-seropositive samples was also unremarkable. Thehistopathologic description of the HIV-1-positivesamples were either unremarkable or the biopsieswere described as having mild, non-specific infl-ammation without evidence of infection. Fig. 1(a)is included to demonstrate the appearance of‘physiologic inflammation’ in the lamina propriawhile Fig. 1(b) identifies the large number ofCD4+ lymphocytes present near the top of onecolonic crypt, nearly all of which are activated,memory cells (Ullrich et al., 1990; Anton et al.,2000). Fig. 1(c) demonstrates the size of forcepsand biopsies used to generate data.

3.2. Optimization of tissue PCR conditions

The LTR sequence was used as our templatefor quantitation studies because it is a requisitefactor for replication and detects both spliced andunspliced HIV-1 RNA (Zack et al., 1990). Inorder to ensure the accurate quantitation of HIV-1 RNA and DNA in mucosal nucleic acid ex-tracts, a series of experiments were conducted toevaluate the efficiency and identify potential in-hibitors of RT-PCR and PCR. For this purpose,a known copy number of HIV-1 LTR RNA wasadded to increasing concentrations (100–1000 ng)of HIV-1-seronegative tissue RNA prior to RT-PCR. Data was only accepted if the R value of


the standard curve was greater than 0.98. A de-crease in specific amplification was observed attissue concentrations in excess of 250 ng. Anupper limit of 250 ng of total tissue RNA wasdefined as optimal for use in the RT-PCR. Similarexperiments were performed using mucosal DNA.Samples were normalized using a �-globin stan-dard, run in parallel as an internal control. Thisenabled comparable amounts of DNA from dif-ferent sized biopsies to be assayed together. Anupper limit of 1000 ng tissue DNA per reactionwas established as optimal for the DNA PCR.These observations could be the result of de-creased efficiency or to the presence of inhibitorsfrom either the tissue itself or the isolationprocess.

3.3. Reco�ery of RNA and DNA from tissue

Having optimized conditions for the down-stream amplification of the tissue HIV-1 LTRregion, a series of experiments were undertaken inwhich exogenous RNA and DNA controls wereused to quantify isolation efficiency. For this pur-

pose, variable amounts of HIV-1 LTR RNA andfirefly luciferase plasmid DNA were added to thesame, HIV-1-seronegative samples prior to ho-mogenization. Recoveries of �89% for RNA and�90% for DNA were defined (Figs. 2 and 3) andshown to be reproducible with minimal intra-as-say variation.

Initial efforts to simultaneously extract RNAand DNA from the same tissue sample had fo-cused on using Trizol only. However, despite be-ing able to recover adequate yields of total RNA,we found the DNA recovery to be low (30–70%)and of poor quality. The described protocol inSection 2, using ULB, permits the simultaneousextraction of RNA and DNA from the sameendoscopic biopsy with high yield for downstreamquantitative RT and PCR reactions under opti-mized conditions.

3.4. Results are reproducible: inter-assay�ariability of tissue PCR

To assess the inter-assay variation followingextraction, mucosal RNA was isolated from 12separate endoscopic biopsies (one biopsy at sixdifferent timepoints obtained from two HIV-1-seropositive patients). The total RNA from all 12samples was subjected to RT-PCR on three sepa-rate days and the mucosal viral load wasquantified (Fig. 4). The results show a high degreeof reproducibility with an inter-assay variabilityof 0.13 log10 (CI: 0.10–0.18; n=79 observations).

3.5. Tissue samples yield inconsistent results whenplasma kits are used to quantitate HIV-1 RNA

Total RNA extracted from tissue samples withdetectable levels of tissue HIV-1 RNA werequantified using the Roche Amplicor® kit forplasma detection of HIV-1. Detected results overa range of dilutions were different from thoseanticipated showing increasing rather than de-creasing copies while the tissue-based assaydemonstrated the expected dilutional effect over anearly 2 log10 range (Fig. 5). The highest concen-tration of tissue RNA produced the highest quan-titation of copies using the tissue assay whereasthis was not the case when using the Amplicor®

Fig. 2. High recovery during extraction from tissue of addedHIV-1 LTR RNA. Increasing amounts of the HIV-1 LTRRNA were added immediately following ULB homogenizationeither pre- or post-extraction to quantitate percent recoveryusing seronegative tissue samples. The gel demonstrates mini-mal difference in recovery between the two routes of addition.The accompanying table shows the percent recovery of gradedamounts of exogenously added HIV-1 LTR RNA when addedpre-extraction.


Fig. 3. High recovery of added luciferase DNA from endoscopic biopsies. Known quantities of luciferase DNA were added to tissuefrom four seronegative controls homogenized in ULB to quantitate recovery from the same samples RNA was to be quantified.Results in table reflect the polyacrylamide gel resolved PCR products and are expressed as percent recovery.

plasma kit. Of note, tissue RNA from subjectc345 was undetectable in all five dilutions persample built into the Amplicor® plasma kit de-spite having detectable levels in the tissue assay.Importantly, co-culture of this sample yielded in-fectious virus (Fig. 7) indicating there was repli-cating HIV-1 present, undetected by Amplicor®

kit but identified in our tissue assay.

3.6. Reproducible detection of mucosal HIV-1RNA and DNA in patients with detectableplasma �iral load

In order to evaluate how representative theHIV-1 RNA and HIV-1 DNA results data fromone biopsy would be with other biopsies col-lected at the same time from the same patient(intra-subject variability), three separate biopsies,

circumferentially, at the 30 cm level in the rec-tosigmoid colon were obtained from patientswith detectable plasma viral load. Biopsies fromsubjects with incompletely suppressed plasma vi-ral load were selected to maximize the likelihoodthat quantifiable levels would be present in thetissue. Table 2 shows the results for HIV-1RNA from three separate biopsies from sevensubjects; results of HIV-1 DNA from the samethree biopsies are shown for three of the sevenpatients.

As a reflection of the representativeness of asingle biopsy, the intra-subject (or inter-biopsy)variability of detected HIV-1 RNA between dif-ferent biopsies was 0.28 log (CI: 0.2–0.42; n=79observations). The variability in the HIV-1 DNAyield between biopsies from the same person wassimilar.


3.7. Reproducible detection of mucosal HIV-1RNA and DNA in subjects with undetectableplasma �iral load

Biopsies and plasma were collected from fivesubjects with reports of undetectable plasma levelson their current antiretroviral regimen. Unde-tectability was defined as �50 copies per ml forthe preceding three months. The purpose of theseexperiments was to assess the sensitivity and effi-ciency of the tissue-based assay in situationswhere plasma viral load is undetectable. Plasmasamples were run using the Amplicor Ultrasen-stive® kit with a lower limit of detection of �50copies of HIV-1 per ml plasma. Results from theassay of mucosal HIV-1 RNA and DNA from thefive subjects are presented in Fig. 6. HIV-1 RNAand DNA are detectable in most situations with

HIV-1 DNA present in �90% of tissue samplestested while HIV-1 RNA is present in �25–40%of tissue samples tested. Tissue HIV-1 RNA isusually detected at lower levels than tissue HIV-1DNA.

3.8. Replication competent HIV-1 is culturedfrom tissue biopsies

Co-culture of mucosal biopsies from six pa-tients with detectable plasma viral was performedto demonstrate that tissue virus quantified usingPCR represents replication competent virus (Fig.7). All samples had measurable levels of mucosalHIV-1 RNA LTR by our PCR assay, includingthe sample (c345) which was read as repeatedlyundetectable when using the Amplicor® kit (Fig.5). These mucosal biopsies resulted in measurable

Fig. 4. Intra-assay variability is low using tissue RNA samples. Twelve separate biopsies were individually extracted and RNA forHIV-1 LTR was amplified on three separate occasions to assess reproducibility. Each result is represented on the chart with themean and S.D. reported in log10 form in the accompanying table; the intra-assay variability is 0.13 log10.


Fig. 5. Plasma assay yields inconsistent results when used for quantitation of tissue HIV-1 RNA. (a) Using Amplicor® kit: extractedmucosal RNA from three subjects known to have quantifiable levels of tissue HIV-1 LTR RNA were diluted in seronegative plasmafor use in the Roche Amplicor® kit. Detected results did not reflect the dilutions with one subject showing increasing copies, anotherwith decreasing copies and one with undetectable copies at all dilutions. Co-culture of mucosal mononuclear cells from subjectc345 (no detectable virus) demonstrated presence of replicating virus. Results are expressed as number of copies per ml. (b) Usingtissue RT-PCR assay: dilutions of mucosal RNA from seropositive samples were quantified by our described methods anddemonstrated the anticipated linear decrease over a nearly 2 log10 range. Tissue-based PCR identified quantifiable levels of virus incases where the Amplicor® kit did not.

amounts of HIV-1 p24 in the co-culture systemwith PHA stimulated PBMCs. Although the num-ber of samples is small, there was a significant,positive correlation between mucosal viral burdenand the amount of p24 produced (R=0.84; P=0.03).

4. Discussion

The predominant site of HIV-1 replication is inthe lymphoid tissue reservoir of the body most ofwhich is located in the gastrointestinal tract. En-

doscopic biopsy of this compartment provides aneasy, accessible means of repeatedly acquiringlymphoid samples for quantifying HIV-1 tissueviral burden. There have been no previous studiesstandardizing and validating the method of ac-quiring mucosal endoscopic biopsies, extractionand PCR amplification of HIV-1 RNA and DNAto demonstrate high recovery, document assayreplicability, and evaluate inter-biopsy variationwithin patients. Although one abstract has re-ported that endoscopic biopsies may yield incon-sistent amounts of HIV-1-infected GALT(Wegner et al., 1997), this has not been generally


the case (Poles et al., 1999; Mitsuyasu et al.,2000). Others have reported detailed quantitativemeasures of HIV-1 RNA load in whole mucosalbiopsy samples but without evidence of themethodological accuracy of the results, especiallywith regard to recovery and representativeness(Fackler et al., 1998; Kotler et al., 1998); with

one exception (Hockett et al., 1999), other re-ports of tissue viral load quantitation are oftenbased on isolated mononuclear cells from themucosa or lymph nodes with or without con-comitant in situ hybridization studies (Wong etal., 1997; Lafeuillade et al., 1998; Dybul et al.,2000).

Table 2Mucosal HIV-1 RNA and DNA are reproducibly detected in concurrently obtained samples from subjects with detectable plasmaviral load

Biopsy c3 Mean S.D.Patient ID no. Biopsy c1 Biopsy c2

HIV RNA copies per �g tissue RNA (log10)345 2.892.30 2.54 0.312.42

0.181.911.97355 2.051.713.87 4.26 3.84 3.99 0.23365

366 0.243.983.824.263.872.852.91 0.102.742.90387

3.51407 3.523.39 0.143.66343 3.103.30 NA 0.282.90

HIV DNA (log10) copies per 2×106 copies �-globin3.83345 3.19 2.94 3.32 0.46

2.802.592.68 0.283.123550.252.923.00 3.12365 2.64

Fig. 6. Mucosal HIV-1 RNA and DNA are detected in mucosal biopsies from subjects with undetectable plasma viral load (�50copies/ml). Total RNA and DNA were concurrently extracted from biopsies from five patients with gel results shown for HIV-1RNA (a) and HIV-1 DNA (b) normalized for �-globin.


Fig. 7. Replicative HIV-1 is detected in co-culture of mucosal biopsies. Mucosal biopsies were co-cultured with activated PBMC,stimulated with PHA for 3 days, and p24 collected at intervals and measured by ELISA. Samples are arranged on the X-axisaccording to increasing mucosal viral RNA LTR detected; p24 measured by ELISA is on the Y-axis. The correlation betweenmucosal viral burden and p24 production is R=0.84 (P=0.03). The accompanying table gives the patient ID numbers, actualcopies of HIV-1 RNA per �g total RNA and p24 in ng/ml as well as the log10-transformed values for each.

In our protocol, time from biopsy acquisitionto freezing in liquid nitrogen is less than 15 s.Experiments designed to assess recovery demon-strated �89% recovery of HIV-1 LTR RNA and�90% recovery of HIV-1 LTR DNA from thesame biopsy sample when added to seronegativetissue samples. The intra-assay variability was0.13 log10 and the inter-biopsy (intra-patient) vari-ability was 0.28 log10. These assays were studiedin populations of patients with detectable andundetectable plasma viral load. The ability toexpress copy number of HIV-1 DNA in terms ofnumber of copies per million cells provides acalculated but more clinically accessible index;HIV-1 RNA remains reported as number ofcopies per �g of total RNA until other optionsevolve to enable the expression of the denomina-tor as ‘per million cells’.

An advantage of the method reported in thispaper is the ability to quantify HIV-1 RNA andDNA from the same biopsy. Due to microscopicheterogeneity in the distribution of mucosallymphocytes, lymphoid aggregates, and follicles,cutting a biopsy in half or using two adjacentsamples for RNA and DNA extraction wouldalways be subject to sampling error, especially incomparative studies of viral RNA and DNA ki-netics under treatment.

Quantitative measurement of mucosal viralload is complicated by the heterogeneous natureof this compartment. These data give assurancethat copy numbers obtained from adjacent biop-sies are reproducible, a crucial feature demon-strating that the results are representative of thatregion without wide biopsy-to-biopsy variation.Although the data demonstrate that single biopsy


results differ little in their HIV-1 RNA or HIV-1DNA content within subjects, in clinical trials thepooling of three separately acquired biopsies mightfurther minimize biological variability. Similarcharacterization of tissue samples obtained else-where in the gut, such as the duodenal mucosa, thesite of infection for many enteric pathogens, re-mains to be assessed. One major advantage of thistechnique is that it uses whole biopsies rather thanisolated T cells and therefore reflects total tissueviral load more accurately.

Attempts to use extracted mucosal RNA inconventional plasma assay kits to quantify HIV-1viral load gave spurious and conflicting results onrepeated assays. Although adaptations of quantita-tive plasma HIV-1 RNA assay have been reported(Cohen et al., 1995; Lafeuillade et al., 1996; Harriset al., 1997; Lafeuillade et al., 1997; Wong et al.,1997; Notermans et al., 1998), translating a solu-tion nucleic acid kit to solid tissue can encounterunforeseen difficulties as shown in this report. Also,the units of quantitation using the Amplicor®

plasma kit are expressed in ‘number of copies perml of plasma’ while the tissue-based assay reportsresults in terms of ‘number of copies per �g totalRNA’. It may still be possible to adapt plasma kitsonce the appropriate amount of RNA to be usedis determined.

Studies of HIV-1 pathogenesis are increasinglyfocused on tissue compartments. It is essential thatsensitive, quantitative and reproducible techniquesfor measuring tissue viral load are developed. Thegut mucosal lymphoid tissue is abundant, easilyaccessible, quickly healing, self-replenishing, anddirectly visible. Endoscopic biopsies are safe, quick,painless, require less than 20 min with minimalpreparation on the part of the patient, and provideaccess to the lymphoid compartment with 100% ofsamples containing lymphocytes. The techniquescan now be used to reliably compare viral dynamicsin mucosal versus blood compartments in primaryinfection as well as following initiation or changesin therapy.

Acknowledgements

Supported by UCLA CFAR Pilot and Feasibil-ity grant (AI28697), UCLA CFAR Mucosal Im-

munology Core (AI28697) and John Boscardin ofthe UCLA CFAR Biostatistics Core (AI28697),NIH grant AI01610 (PAA), NIH grant AI01668(MP), Macy’s Foundation and the OppenheimerBrothers Foundation. We would like to thank Dr.Kathie Grovit-Ferbas for providing the LTR RNAstandards.

References

al-Mulla, W., Church, D., Gill, M.J., 1997. Phenotypic vari-ations and switches in HIV isolated from the blood andthe gastrointestinal tissues of patients with HIV-1 infec-tion. J. Med. Virol. 52, 31–34.

Andersson, J., Fehniger, T.E., Patterson, B.K., Pottage, J.,Agnoli, M., Jones, P., Behbahani, H., Landay, A., 1998.Early reduction of immune activation in lymphoid tissuefollowing highly active HIV therapy. AIDS 12, F123–F129.

Anton, P.A., Elliot, J., Poles, M.A., McGowan, I.M.,Matud, J., Hultin, L.E., Grovit-Ferbas, K., Mackay,C.R., Chen, I.S.Y., Giorgi, J.V., 2000. Enhanced levels offunctional HIV-1 co-receptors on human mucosal T cellsdemonstrated using intestinal biopsy tissue. AIDS 14,1761–1765.

Cavert, W., 1998. In vivo detection and quantitation of HIVin blood and tissues. AIDS 12 (Suppl. A), S27–S34.

Cavert, W., Notermans, D.W., Staskus, K., Wietgrefe, S.W.,Zupanic, M., Gebhard, K., Henry, K., Zhang, Z.Q.,Mills, R., McDade, H., Goudsmit, J., Danner, S.A.,Haase, A.T., 1997. Kinetics of response in lymphoid tis-sues to antiretroviral therapy of HIV-1 infection. Science276, 960–964.

Cerf-Bensussan, N., Guy-Grand, D., 1991. Intestinal intraep-ithelial lymphocytes. Gastroenterol. Clin. North Am. 20,549–576.

Cohen, O.J., Pantaleo, G., Holodniy, M., 1995. Decreasedhuman immunodeficiency virus type 1 plasma viremiaduring antiretroviral therapy reflects downregulation ofviral replication in lymphoid tissue. Proc. Natl. Acad.Sci. USA 92, 6017–6021.

Cohen, O.J., Pantaleo, G., Lam, G.K., Fauci, A.S., 1997.Studies on lymphoid tissue from HIV-infected individu-als: implications for the design of therapeutic strategies.Springer Semin. Immunopathol. 18, 305–322.

Dybul, M., Chun, T.-W., Ward, D.J., Hertogs, K., Larder,B., Fox, C.H., Orenstein, J.M., Baird, B.F., Li, Y.,Green, L.G., Engel, D., Liu, S., Mican, J.M., Fauci,A.S., 2000. Evaluation of lymph node virus burden inhuman immunodeficiency virus-infected patients receivingefavirenz-based protease inhibitor-sparing highly activeantiretroviral therapy. J. Infect. Dis. 181, 1273–1279.

Fackler, O.T., Schafer, M., Schmidt, W., Zippel, T., Heise,W., Schneider, T., Zeitz, M., Riecken, E.-O., Mueller-


Lantzsch, N., Ullrich, R., 1998. HIV-1 p24 but not provi-ral load is increased in the intestinal mucosa comparedwith the peripheral blood in HIV-infected patients. AIDS12, 139–146.

Fantini, J., Cook, D.G., Nathanson, N., Spitalnik, S.L., Gon-zalez-Scarano, F., 1993. Infection of colonic epithelial celllines by type 1 human immunodeficiency virus is associatedwith cell surface expression of galactosylceramide, a poten-tial alternative gp120 receptor. Proc. Natl. Acad. Sci. USA90, 2700–2704.

Geijtenbeek, T.B., Kwon, D.S., Torensman, R., van Vliet, S.J.,van Duijnhoven, G.C., Middle, J., Cornelissen, I.L., Not-tet, H.S., KewalRamani, N.V., van Kooyk, Y., 2000. DC-SIGN, a dendritic cell-specific HIV-1-binding protein thatenhances trans-infection of T cells. Cell 100, 587–597.

Haase, A.T., Henry, K., Zupancic, M., Sedgewick, G., Faust,R.A., Melroe, H., Cavert, W., Gebhard, K., Staskus, K.,Zhang, Z.Q., Dailey, P.J., Balfour, H.H. Jr., Erice, A.,Parelson, A.S., 1996. Quantitative image analysis of HIV-1infection in lymphoid tissue. Science 274, 985–999.

Harris, M., Patenaude, P., Cooperberg, P., Filipenko, D.,Thorne, A., Raboud, J., Rae, S., Dailey, P., Chernoff, D.,Todd, J., Conway, B., Montaner, J.S.G., 1997. Correlationof virus load in plasma and lymph node tissue in humanimmunodeficiency virus infection. J. Infect. Dis. 176,1388–1392.

Hockett, R.D., Kilby, J.M., Derdeyn, C.A., Saag, M.S.,Sillers, M., Squires, K., Chiz, S., Nowak, M.A., Shaw,G.M., Bucy, R.P., 1999. Constant mean viral copy numberper infected cell in tissues regardless of high, low, orundetectable plasma HIV RNA. J. Exp. Med. 189, 1545–1554.

Kewenig, S., Schneider, T., Hohloch, K., Lampe-Dreyer, K.,Ullrich, R., Stolte, N., Stahl-Hennig, C., Kaup, F.J., Stall-mach, A., Zeitz, M., 1999. Rapid mucosal CD4+ T-celldepletion and enteropathy in simian immunodeficiencyvirus-infected rhesus macaques. Gastroenterology 116,1115–1123.

Kotler, D.P., Shimada, T., Snow, G., Winson, G., Chen, W.,Zhao, M., Inada, Y., Clayton, F., 1998. Effect of combina-tion antiretroviral therapy upon rectal mucosal HIV RNAburden and mononuclear cell apoptosis. AIDS 12, 597–604.

Lafeuillade, A., Chollet, L., Hittinger, G., Profizi, N., Costes,O., Poggi, C., 1998. Residual human immunodeficiencyvirus type 1 RNA in lymphoid tissue of patients withsustained plasma RNA of �200 copies/ml. J. Infect. Dis.177, 235–238.

Lafeuillade, A., Poggi, C., Profizi, N., Tamalet, C., Costes, O.,1996. Human immunodeficiency virus type 1 kinetics inlymph nodes compared with plasma. J. Infect. Dis. 174,404–407.

Lafeuillade, A., Poggi, C., Tamalet, C., Profizi, N., Tourres,C., Costes, O., 1997. Effects of a combination of zi-dovudine, didanosine, and lamivudine on primary humanimmunodeficiency virus type 1 infection. J. Infect. Dis. 175,1051–1055.

Lapenta, C., Boirivant, M., Marini, M., Santini, S.M.,Logozzi, M., Viora, M., Belardelli, F., Fais, F., 1999.Human intestinal lamina propria lymphocytes are natu-rally permissive to HIV-1 infection. Eur. J. Immunol. 29,1202–1208.

Mitsuyasu, R.T., Anton, P.A., Deeks, S.G., Scadden, D.T.,Connick, E., Downs, M.T., Bakker, A., Roberts, M.R.,June, C.H., Jalali, S., Lin, A.A., Pennathur-Das, R., Hege,K.M., 2000. Prolonged survival and tissue trafficking fol-lowing adoptive transfer of CD4zeta gene-modified au-tologous CD4+ and CD8+ T cells in humanimmunodeficiency virus-infected subjects. Blood 96, 785–793.

Mowat, A.M., Viney, J.L., 1997. The anatomical basis ofintestinal immunity. Immunol. Rev. 156, 145–166.

Notermans, D.W., Jurriaans, S., de Wolf, F., Foudraine,N.A., de Jong, J.J., Cavert, W., Schuwirth, C.M., Kauff-mann, R.H., Meenhorst, P.L., McDade, H., Goodwin, C.,Leonard, J.M., Goudsmit, J., Danner, S.A., 1998. Decreaseof HIV-1 RNA levels in lymphoid tissue and peripheralblood during treatment with ritonavir, lamivudine andzidovudine. AIDS 12, 167–173.

Pantaleo, G., Graziosi, C., Butini, L., Pizzo, P.A., Schnittman,S.M., Kotler, D.P., Fauci, A.S., 1991. Lymphoid organsfunction as major reservoirs for human immunodeficiencyvirus. Proc. Natl. Acad. Sci. USA 88, 9838–9842.

Perrin, L., Yerly, S., Marchal, F., Schockmel, G.A., Hirschel,B., Fox, C.H., Pantaleo, G., 1998. Virus burden in lymphnodes and blood of subjects with primary human im-munodeficiency virus type 1 infection on bitherapy. J.Infect. Dis. 177, 1497–1501.

Poles, M., Elliot, J., Brown, S., Shi, D.-P., Chiong, S., Hege,K.M., McGowan, I.M., Lenz, H., Chen, I.S.Y., Anton,P.A., 1999. HIV-1 is detectable in mucosal biopsies inpatients with undetectable plasma viral loads. In: SixthConference on Retroviruses and Opportunistic Infections,Chicago, IL, January 1999 (abstract 160).

Poles, M.A., Elliott, J., Vingerhoets, J., Michiels, L., Schol-liers, A., Bloor, S., Larder, B., Hertogs, K., Anton, P.A.,2001. Concordance in Genotypic and Phenotypic Resis-tance Profiles of HIV-1 RNA Isolated From Gastrointesti-nal Mucosal Biopsies, PBMCs and Plasma. J. Infect. Dis.183, 143–148.

Sei, S., Kleiner, D.E., Kopp, J.B., Chandra, R., Klotman,P.E., Yarchoan, R., Pizzo, P.A., Mitsuya, H., 1994. Quan-titative analysis of viral burden in tissues from adults andchildren with symptomatic human immunodeficiency virustype 1 infection assessed by polymerase chain reaction. J.Infect. Dis. 170, 325–333.

Ullrich, R., Schieferdecker, H.L., Ziegler, K., Riecken, E.-O.,Zeitz, M., 1990. Gamma delta T cells in the human intes-tine express surface markers of activation and are preferen-tially located in the epithelium. Cell Immunol. 128,619–627.

Veazey, R.S., DeMaria, M., Chalifoux, L.V., Shvetz, D.E.,Pauley, D.R., Knight, H.L., Rosenweig, M., Johnson,


R.P., Desrosiers, R.C., Lackner, A.A., 1998. Gastrointesti-nal tract as a major site of CD4+ T cell depletion andviral replication in SIV infection. Science 280, 427–431.

Wegner, S., Anderson, D., Denobile, J., Cotelingam, J., Nau,M., Vahey, M., Mayers, D., Birx, D., Frankel, S., 1997.Rectal and lymph node (LN) biopsies in early stage HIV-1infected patients. In: Fourth Conference on Retrovirusesand Opportunistic Infections, Washington, DC, January1997 (abstract 239)

Wong, J.K., Gunthard, H.F., Havlir, D.V., Zhang, Z.Q.,Haase, A.T., Ignacio, C.C., Kwok, S., Emini, E., Richman,D.D., 1997. Reduction of HIV-1 in blood and lymph nodes

following potent antiretroviral therapy and the virologiccorrelates of treatment failure. Proc. Natl. Acad. Sci. USA94, 12574–12579.

Zack, J.A., Arrigo, S.J., Weitsman, S.R., Go, A.S., Haislip,A., Chen, I.S.Y., 1990. HIV-1 entry into quiescent primarylymphocytes: molecular analysis reveals a labile, latentviral structure. Cell 61, 213–222.

Zeitz, M., Greene, W.C., Peffer, N.J., James, S.P., 1988.Lymphocytes isolated from the intestinal lamina propria ofnormal nonhuman primates have increased expression ofgenes associated with T cells activation. Gastroenterology94, 647–655.

.

Documents

Sensitive and reproducible quantitation of mucosal HIV-1 ... · Journal of Virological Methods 95 (2001) 65–79 Sensitive and reproducible quantitation of mucosal HIV-1 RNA and DNA