bio markers of cervicovaginal inflammation for the.4[1]

8/8/2019 Bio Markers of Cervicovaginal Inflammation for the.4[1]

http://slidepdf.com/reader/full/bio-markers-of-cervicovaginal-inflammation-for-the41 1/8

REVIEW

Biomarkers of Cervicovaginal Inflammation for the Assessment of Microbicide Safety

James E. Cummins, Jr., PHD,* and Gustavo F. Doncel, MD, PHD†

Abstract: The human cervicovaginal mucosa is the primary target of

HIV-1 infection during male to female transmission. This tissue contains

the the full spectrum of cell types and immune modulators that comprise

both the innate and adaptive arms of the immune system. Mounting

evidence indicates that mucosal epithelial cells are sentinels of the female

reproductive tract, producing innate immune mediators that control the

vaginal microflora under normal conditions. Recent studies, however,

indicate that certain factors secreted in response to another pathogen or

after exposure to a vaginal product may in fact enhance infection by

HIV-1. Mucosal inflammation and CD4 cell activation as well as disrup-

tion of TLR function and epithelial integrity represent potential causes forsuch effect. It is therefore important to make sure that vaginal products,

including microbicides, do not disrupt the structure or function of the

cervicovaginal mucosa. Although a number of biomarkers have been

linked to microbicide-induced cervicovaginal inflammation and many of

these markers have been measured in preclinical and clinical assays, there

are currently no data that demonstrate a correlation between any one

marker and susceptibility to HIV-1 infection in humans. To date, the lack

of a validated biomarker of cervicovaginal safety represents a gap in the

knowledge base that hinders the rational and expeditious selection of

microbicide candidates entering clinical trials. Currrent discovery efforts

and preclinical assessment of microbicide safety use an integrated sequen-

tial evaluation system that includes cell-based models, explant-based mod-

els, and animal-based models. Relevant research in these areas is yielding

new assays and biomarkers that, if validated, will be essential to the

rational selection of microbicide candidates for efficacy trials.

There is general agreement among scientists that identifica-tion of biomarkers predictive of microbicide safety and

efficacy is urgently needed, yet no clear candidates are ready tobe implemented in the preclinical selection of microbicides ortheir evaluation in clinical trials. Lack of validation is the mostimportant hurdle, but not the only one. Although a number of potential biomarkers, such as cytokines, chemokines, inflam-matory, and innate factors, have been identified, the influenceof physiological and pathologic events such as ejaculation,menstrual cycle, pregnancy, and presence of altered microfloraor pathogens on their expression has not been clearly eluci-

dated. As a result, the levels of these biomarkers in the genitaltract are variable, and such variability confounds the interpre-tation of results derived from mucosal exposure to microbi-cides. In addition to an improved understanding of the eventscontrolling innate and adaptive immunity, both cellular andmolecular mechanisms of mucosal inflammation should beconsidered in choosing robust biomarker candidates. This isparticularly important given the complexity and cost associatedwith implementing additional biomarker sampling in clinicaltrials. This manuscript briefly reviews some of the basic con-

cepts underlying the preclinical development and clinical im-plementation of current models and biomarkers of microbicidesafety. The review is mostly based on the presentations of Drs.John Fahey (Dartmouth University), Bruce Horwitz (HarvardUniversity), Gustavo Doncel (CONRAD-Eastern VirginiaMedical School), Radiana Trifonova (Harvard University),Betsy Herold (Mount Sinai Medical School), and Sharon Hill-ier (University of Pittsburgh), and the subsequent discussionsled by Drs. James Cummins (Southern Research Institute),Carolyn Deal (National Institutes of Health), Deborah Ander-son (Boston University), and Ken Mayer (Brown University)during a session called “Biomarkers of cervicovaginal inflam-mation” held at the conference entitled “Biomarkers for eval-uation of vaginal microbicides and contraceptives: Discovery

and early validation,” organized by CONRAD and the Alliancefor Microbicide Development in November of 2006. We haveused these presentations as a starting point for this review,updating the topics whenever significant progress has occurred.

MUCOSAL IMMUNITY IN THE FEMALE GENITALTRACT

The female reproductive tract contains the full spectrumof cell types and immune modulators that comprise both theinnate and adaptive arms of the immune system. As part of thefirst line of defense, cells of the innate immune system (epithelialcells, macrophages, dendritic cells, neutrophils, and natural killercells) mediate antigen presentation, phagocytosis, and direct kill-ing of pathogens. In addition, these cell types produce solubleimmune modulators such as cytokines, chemokines, and antimi-crobial peptides (e.g., defensins and secretory leukocyte proteaseinhibitor [SLPI]). As part of the adaptive immune system, B andT cells respond to antigens present in the mucosa and in turnmount antibody or cell-mediated immune responses. Leukocytescan be found in both the upper and lower female reproductivetract, and fluctuations in levels of lymphocytes and monocytesoccur across the menstrual cycle.1 In addition, lymphoid aggre-gates present in the uterine endometrium represent immune sitescapable of generating both B and T cell immune responses in closeassociation with the uterine epithelium.2

Mounting evidence indicates that mucosal epithelial cellsare sentinels of the female reproductive tract with broad activity

against bacteria, viruses, and fungi. Enzymatic digestion of uterinetissue yields primary uterine epithelial cells that have been used toexamine innate immune responses.3 Once cultured ex vivo, these

From the *Southern Research Institute, Frederick, Maryland; and†CONRAD, Eastern Virginia Medical School, Norfolk, Virginia

The authors thank for the essential contribution of the speakers of thissession through their presentations to the information presented inthis review; and USAID and The Bill and Melinda Gates Founda-tion for their support.

Supported by USAID and The Bill and Melinda Gates Foundation.The views expressed by the authors do not necessarily reflect those of

the funding agenciesCorrespondence: Gustavo F. Doncel, MD, Eastern Virginia Medical

School, 601 Colley Avenue, Norfolk, VA 23507. E-mail:[email protected].

Received for publication July 6, 2008, and accepted December 5, 2008.

DOI: 10.1097/OLQ.0b013e3181994191Copyright © 2009 American Sexually Transmitted Diseases AssociationAll rights reserved.

S84 Sexually Transmitted Diseases ● Volume 36, Number 3, Supplement March 2009



cells express toll-like receptors (TLR’s) 1 to 9 and the -defensins1 to 3. Stimulation of these cells with the TLR3 agonist Poly (I:C)induces the expression of cytokines/chemokines (tumor necrosisfactor (TNF-), interleukin [IL]-6, granulocyte acrophagecolonystimulating factor (GM-CSF), monocyte chemotactic protein(MCP-1), and IL-8), -defensins, and interferon-. Reproductive

hormones may modify these responses. Estradiol, for instance,acts directly at multiple levels in epithelial cells and can increaseexpression of the innate factors SLPI and human -defensin(HBD)1. Estradiol inhibits IL-1-stimulated secretion of HBD2and IL-8 by downregulating the IL-1 receptor on uterine epithelialcells,4 yet it augments lipopolysaccharide (LPS)-induced IL-1production in uterine macrophages.5 In addition, antibacterial ac-tivity in apical secretions from uterine epithelial cells is mediatedby SLPI, and estradiol can enhance both antibacterial activity andSLPI levels.6

Although less characterized than uterine cells, vaginalepithelial cells are also responsive to reproductive hormonesand microbial antigens, producing immune innate mediatorsthat under normal conditions control the vaginal microflora. In

a very recent article, certain polyanionic compounds wereshown to interfere with TLR ligand mediated innate responsesby human female reproductive tract cells, highlighting the needto study the effects of microbicides on mucosal immune mech-anisms of defense.7

Although the secretion of innate factors by mucosalepithelium is generally thought to provide protective benefitagainst infection with pathogens, recent studies indicate thatcertain factors secreted in response to one pathogen may in factenhance infection by another.8 In a cervicovaginal tissue cul-ture system, expression of human defensin (HD)5 and HD6 wasinduced in response to Neisseria gonorrhoeae (GC) infection,and conditioned medium from GC-exposed cervicovaginal ep-ithelial cells with elevated levels of HD5 enhanced HIV infec-

tion in primary T cells.

CELLULAR AND MOLECULAR MECHANISMS OFMUCOSAL INFLAMMATION

While microflora can drive inflammation in the genitalmucosa, additional evidence indicates that the same is true inother mucosal sites such as the intestine. Two inflammatorybowel diseases, Crohn’s disease and ulcerative colitis (UC), areinfluenced by 3 etiologic factors: genetic susceptibility, micro-flora challenge, and immune dysregulation. Although each dis-ease is characterized by a different TH1 (Crohn’s) or TH2 (UC)cytokine profile, in both cases, a failure to downregulate theimmune response leads to disease. In mouse disease models,

mice with a disrupted IL-2 gene develop an UC-like disease,9and resident enteric bacteria were able to induce colitis inIL-10-deficient mice.10,11 Although both TH1 and TH17 effec-tor T cells are involved in driving mucosal inflammation, IL-10production by regulatory T cells is required to inhibit it.12–14

IL-10-deficient mice lacking MyD88, an adapter protein in theTLR pathway, are resistant to colitis.15 Polymorphisms in otherintracellular pattern recognition receptors, such as Nod2, canincrease mucosal inflammation by activating the nuclear factor B (NF-B) pathway in macrophages or reducing defensinproduction in intestinal epithelial cells.

Similar mechanisms are operational in genital mucosalinflammation. Leukocyte traffic and activation are mediated byproinflammatory cytokines and chemokines, e.g., IL-1, IL-6,

and IL-8, which have been detected in vaginal secretions inassociation with epithelial damage and infections. In vitro andanimal studies have shown changes in mucosal TH1-associated

cytokines in the presence of Chlamydia trachomatis infectionand with its progression to the upper genital tract or clear-ance.16 Increased IL-12 and decreased IL-2, observed com-monly during mucosal inflammation, are important features of mucosal immune defense against C. trachomatis infection.

TLR have recently been identified as fundamental compo-

nents of the innate immune response to bacterial pathogens in thelower female genital tract. This site provides a first line of defenseagainst microbial pathogens, although remaining tolerant to acomplex biosystem of resident microbiota. Epithelial cells derivedfrom normal human vagina, ectocervix, and endocervix expressmRNA for TLR1, 2, 3, 5, and 6. However, they fail toexpress TLR4 and MD2, 2 essential components of the receptorcomplex for LPS in phagocytes and endothelial cells.17 Consistentwith this, endocervical epithelial cells are unresponsive to protein-free preparations of LPS from N. gonorrhoeae and Escherichiacoli. However, they are capable of responding to whole Gram-negative bacteria and bacterial lysates, as demonstrated by NF-Bactivation and proinflammatory cytokine production.

A number of recent publications further indicate the role

of TLR’s in innate immunity and highlight changes in thegenital tract that may influence the protective role of thesereceptors. The menstrual cycle can influence expression of theTLR’s,18,19 TLR ligand/agonists can effect protection againstherpes-simplex virus (HSV)-2 infection.20 As indicated above,the need to characterize the impact of microbicide candidateson genital epithelial TLR mediated responses seems warranted.

MODELS AND BIOMARKERS FOR PRECLINICALEVALUATION OF MICROBICIDE SAFETYThe efficacy of a microbicide depends on the balance

between its specific activity and safety. The ideal microbicidewould retain its activity, while not disrupting the integrity of

the epithelial barrier, not inducing mucosal inflammation, andnot interfering with innate immune responses. Pathologic con-ditions associated with increased cervicovaginal transmissionof HIV include sexually transmitted infections,21,22 vagini-tis,23,24 disturbances of the vaginal flora,25,26 and lesions of cervical/vaginal mucosa.27 As previously described, the hypo-thetical mechanism of microbicide-induced mucosal toxicityand enhancement of HIV infection is based on the “sensing”ability of the cervicovaginal epithelium and includes (1) cellinjury leading to disruption of the epithelial barrier, (2) cellactivation with release of soluble mediators of inflammation,and (3) cell inhibition that interferes with the delicate mecha-nisms of innate immunity.28 (Figure 1)

The current preclinical assessment of microbicide mu-

cosal safety uses an integrated sequential evaluation systemthat includes cell-based models, explant-based models, andanimal-based models (Table 1). Cell-based models are useful asa first-line screen to identify candidates that are significantlycytotoxic. They are also useful for mechanistic studies andidentification of potential biomarkers. Pooled data from 5 lab-oratories indicated a striking consistency for time-dependentnonoxynol-9 (N-9) cytotoxicity across the laboratories and celltypes.29

In addition to cell viability and epithelial barrier function(permeability), other toxicity endpoints used in cell-based mod-els include measurement of cytokines, chemokines, inflamma-tory mediators, innate immunity mediators, transcription fac-tors, and other molecules (Table 1). These other toxicity

endpoints are not as consistent as cytotoxicity, since the dataare challenging to analyze because of variability in the analyteconcentration and the system used to measure it.30 In search of

Biomarkers of Cervicovaginal Inflammation

Sexually Transmitted Diseases ● Volume 36, Number 3, Supplement March 2009 S85



data that are more consistent and easier to interpret, investiga-tors have resorted to the assessment of general mucosal re-sponses utilizing approaches that range from genomic analysisof epithelial cells in contact with microbicides to the evaluationof epithelial permeability to HIV. As examples of these ap-proaches, recent data show increased intracellular expression of COX-2 and PGE2 inflammatory mediators in genital epithelialcells incubated with N-9 (Zalenskaya and Doncel, unpub-lished).31 Also from recent data, cellulose sulfate appears todecrease transepithelial resistance of uterine cells, increasing itspermeability to HIV and the infection of target cells in adual-chamber model.32

Explant-based models for assessment of microbicidesinclude a variety of tissues (cervical, vaginal, foreskin, penile,

colorectal, and tonsilar) and endpoints.33–39 Although there is aprogressive loss of tissue morphology in these explant tis-sues,33,34 strengths of the explant models include the range of cell targets present in mucosal tissues, better tolerance to for-mulations, and parallel efficacy testing. Weaknesses includelimited number of tissues, high variability, need for Institu-tional Review Board approval, and technical demands. Com-mercially available reconstructed epithelia from the humangenital tract are becoming increasingly popular because theylack some of these problems.40 Among their advantages areincreased amount of tissue available, reduced variability, andfewer regulatory constraints; however, a limitation is that anexogenous source of immune target cells must be added to thesystem. Altogether, these models offer a way to test candidate

InnateImmunity

CellActivation

Cell Injury andDysfunction(possible death)

Inflammation

MucosalInfection

Microbicide SolubleMediators

Disruptionof EpithelialBarrier

MucosalInfection

Cell Inhibition

HIVinfection

Figure 1. Hypothetical model of microbicide-induced cervocovagi-nal mucosal alteration and possi-ble enhancement of HIV infection.Through its interaction with thecervicovaginal epithelium and its

associated microbial biofilm, a mi-crobicide may disrupt epithelialintegrity and permeability, inducerelease of proinflammatory medi-ators, and/or alter mucosal innateimmunity. Any of these phenom-ena could lead, directly or indi-rectly, to enhanced HIV infection.Modified with permission fromDoncel et al., 2004.

TABLE 1. Preclinical Assessment of Microbicide Mucosal Safety Through an Integrated Sequential Evaluation System

Assays Model Endpoints

Cell-based(to test active pharmaceutical ingredientsAPIs)

Vaginalectocervical

EndocervicalEndometrialIntestinalImmune

Viability (cytotoxicity), permeability (permeation,electrical resistance, tight junctions)

Cytokines (e.g., IL-1, IL-6, TNF-)Chemokines (e.g., IL-8, macrophage inflammatory

protein (MIP), RANTES)Inflammatory mediators (e.g., prostaglandins (PGs),

vascular endothelial growth factor (VEGF),macrophage inflammatory protein (MPO))

Innate immunity mediators (e.g., defensins, SLPI,lactoferrin (Lf), gp340)

Transcription factors (e.g., NF-B, AP-1)Others (lgG, IL-1RA, IP-10)

Explant-based (to test APIs and formulations) Cervical Same as aboveColorectal

foreskinLymphoid

Animal-based (to test formulations) RabbitMouse

HistopathologyImmune infiltrates (phenotype and activation status)

in tissues and CVLs

Monkey Soluble markers (e.g., inflammatory, innate, vascular)Antiviral/antibacterial activity of CV secretionsMucosal integrityMicroflora

CV indicates cervicovaginal.[1]Used with permission from Gustavo Doncel.

Cummins JR et al.




products in a system with cell types specific to the mucosa,unlike traditional screening methods that rely or peripheralblood mononuclear cells or cell lines. Thus, explant-basedmodels provide a “step-up” screening, critical for assessment of preliminary formulations, allowing evaluation of nonepithelialcomponents and tissue-based microbicide antiviral activity, and

immune mediator induction.35,41Animal-based models for assessment of microbicides

include three primary animals (rabbit, mouse, and monkey)and several endpoints (histopathology, leukocyte infiltrationand phenotype, cytokines and soluble markers, epithelialdamage, apoptosis, colposcopy, and microflora). Strengthsof the animal models include the fact that they allow for theevaluation of the whole-organ response to stimuli (e.g.,microbicides), the existence of dynamic responses, and apossible systemic component of a local (vaginal) immunoin-flammatory reaction. They are also the preferred model toassess full-strength formulations. Weaknesses include issuesrelated to animal welfare, Institutional Animal Care and UseCommittee approval, nonhuman tissues, labor intensity, and

expense. These models are utilized by some microbicidedevelopers as “gate keepers” to Phase I clinical trials. How-ever, from a regulatory point of view, the US Food and DrugAdministration only recommends the rabbit vaginal irrita-tion (RVI) model for nonclinical safety evaluation.42

This model has been improved and expanded to includesoluble and cellular markers of inflammation and vascularactivation.28 Rabbits are treated with 1 ml of full-strengthformulation, intravaginally, once a day for 3 to 14 days.Cervicovaginal lavages are collected at baseline, during thetreatment period and 24 and 48 hours after the last dose.Cells are separated from the supernatants and immunophe-notyped, and supernatants are evaluated for soluble markersof inflammation, vascular permeability, and epithelial dis-

ruption. Cervicovaginal tissues are collected 48 hours afterthe last dose and evaluated for epithelial disruption, immuneinfiltrates, and markers of enhanced vascular permeability(see below). For irritating compounds such as N-9, themodel is sensitive enough to detect mucosal alterations afteronly 3 daily doses.43 The model can also be extended for 10to 14 days to cover the duration of the initial phase I safetytrials, however, only the standard RVI is required for regu-latory purposes. Increased IL-1 levels were observed inthis model after application of N-9 and other surface-activeagents, and these levels correlated well with the standardRVI histopathologic score. Characterization of the mucosalinfiltrates revealed an influx of activated CD4 cells into thevaginal lumen and tissues, which was not limited to the most

irritating agents, i.e., N-9 and benzalkonium chloride (BZK).Even compounds that apparently did not affect mucosal integrityinduced lymphocyte recruitment and activation.28 Because im-mune cell activation facilitates HIV infection and replication,44

it is extremely important to test microbicide candidates for theirability to trigger this event, especially after prolonged repeatedexposure to the cervicovaginal mucosa.

BIOMARKERS OF VASCULAR PERMEABILITYAND LEUKOCYTE INFILTRATION IN THE

VAGINAL MUCOSAAs described above, epithelial damage or activation in

the mucosa leads to IL-1 secretion and transactivation of

NF-B, resulting in release of cytokines (IL-1, IL-6, andTNF-) and chemokines (IL-8, IL-10, and macrophage inflam-matory protein (MIP) 3). The resulting influx of neutrophils

and HIV-1 target cells can lead to subsequent HIV-1 infectionand virus shedding. Although concentrations of interleukins invaginal secretions correlate with epithelial damage and inflam-mation in vivo,43,45,46 these cytokines/chemokines are not leu-kocyte specific, have pleiotropic and opposing functions, andcan be blocked by endogenous antagonists. Thus additional

biomarkers, such as those specific for vascular change andleukocyte activation, are needed to facilitate the assessment of HIV-1 transmission risk.

At sites of inflammation, specific receptors on leuko-cytes and endothelial cells are involved in leukocyte adhesionto endothelium and subsequent extravasation of these cells intomucosal tissues. E-selectin, expressed by cytokine-activatedendothelial cells and CD4 T cells, mediates endothelial ad-hesion of polymorphonuclear leukocytes and monocytes, earlyin the process of leukocyte extravasation.47,48 Although thesoluble form is present in serum and inflamed synovial fluid, itspresence and function in vaginal fluids has not been character-ized. Vascular cell adhesion molecule-1 (VCAM-1), expressedby endothelial cells, macrophages, fibroblasts, and squamous

epithelial cells, mediates monocyte adhesion and the late eventsof extravasation.49 The soluble form circulates in serum and iselevated in synovial fluid, cerebrospinal fluid, and bronchoal-veolar lavage during inflammation. Again, soluble VCAM-1has not been characterized in the vaginal milieu. The solubleform of CD14, expressed by monocytes, macrophages, andneutrophils, has been correlated to HIV-1 shedding in thefemale genital tract.50 Myeloperoxidase (MPO) is an activationmarker specific for neutrophils and responsible for the antibac-terial properties of polymorphonuclear leukocytes. Both CD14and MPO are leukocyte specific and are not expressed byepithelial cells.

In the improved RVI model, formulated microbicidescan be evaluated for their effects in tissues by analysis of

histopathology, cellular infiltrate, apoptosis, and mucosalpermeability markers.28 In cervicovaginal lavages, superna-tants can be evaluated for cytokines and leukocyte markers,and the cell pellet can be used to quantitate, phenotype, andmeasure the activation status of leukocytes. When BZK,N-9, and sodium dodecyl sulfate were tested in this model,leukocyte infiltration correlated with activation of NF-Band activator protein (AP)-1 in the rabbit mucosal epitheli-um.51 In lavage samples, these compounds induced solublemarkers (E-selectin, VCAM-1, and MPO) that paralleled thetissue data (Figure 2). Because these new vascular andinflammatory markers can also be detected in human cervi-covaginal lavages, they should be considered as potentialbiomarker candidates for the selection of vaginal microbi-

cides with favorable safety profiles.

DEVELOPMENT OF A MOUSE MODEL ANDINNATE IMMUNITY MARKERS PREDICTIVE OF

MICROBICIDE SAFETYThere are a number of limitations for the current pre-

clinical and clinical microbicide safety trials. While preclinicalstudies focus on cytotoxicity in cell lines and explants, anacceptable selectivity index for vaginally applied drugs is stillnot known. Although preclinical studies rely on the standardRVI model and clinical trials rely on colposcopy and adverseevents, past experience with N-9 suggests that these strategiesalone may not be sufficient to predict increased risk of HIV-1

acquisition. Current data indicate that measurement of cyto-kines and other inflammatory markers may provide insight intothe potential cytotoxicity of surfactant agents; however, the





biologic or functional significance of increased cytokine levelsin vaginal secretions is not known because normal physiolog-ical ranges are not clearly defined.

The goal is to identify or develop assays that predict safetyof products that will be used repeatedly and intermittently, bothvaginally and rectally. The ideal microbicide would: (1) have nocytotoxicity, resulting in a high selectivity index, (2) be nonin-flammatory, (3) have no deleterious effect on normal vaginal flora,(4) preserve or enhance mucosal immunity, (5) have little or nosystemic absorption, and (6) have little or no selection for resistantvariants.

As a complement to the RVI models and colposcopy/ cytokine analysis in Phase I clinical trials, a comprehensivemouse model is available that is well-characterized immuno-logically and is capable of being infected with a mucosalpathogen (HSV-2).52–54 This model can be used to evaluate theimpact of frequent or intermittent topical microbicide applica-tion on inflammatory responses and mucosal immunity. Moreimportantly, the biologic significance of any observed changesin immune mediators can be related to enhanced susceptibilityto HSV infection.

In this mouse model, treatment with N-9 resulted insignificant increases in monocyte chemotactic proteing(MCP)-1, MIP-2, and IL-1 in vaginal washes at days 3 and 7compared to baseline, and PRO 2000 treatment resulted in

significant increases in MIP-2 levels on day 3. There was asignificant increase in the number of CD45 leukocytes invaginal tissue after 7 days of treatment with N-9 that was not

observed after PRO 2000 treatment. In addition, N-9 treatment,but not PRO 2000, induced epithelial cell disruption and in-flammation, and triggered intracellular NF-B and AP-1. In afunctional study, mice were treated intravaginally with themicrobicides for 7 days, and 12 hours after the last dose themice were challenged with low dose HSV-2. Although N-9significantly increased susceptibility to HSV, PRO 2000showed no increase in susceptibility.55

Mounting evidence indicates the presence of innate im-mune factors in the genital mucosa that act as protective media-tors. Cervical secretions from healthy women can protect human

epithelial cells against in vitro infection with HSV56 and HIV.57Cationic peptides (human neutrophil peptides and other defensins)in vaginal secretions have been identified that contribute to intrin-sic antiviral activity.58 Although results from a recent clinical trialwith PRO 2000 indicate a decline in select immune mediators(IL-6, IL-8, HBD-2, and SLPI), there was no loss in intrinsicantiviral or antibacterial activity in the cervicovaginal samples.59,60

These results warrant long-term studies to determine whether asustained loss of immune mediators can lead to increased suscep-tibility to infection. Testing of additional compounds is necessaryto assess whether these assays are predictive of clinical safety andwhether innate immune mediators should be included in currenttesting algorithms.

Because validated preclinical biomarkers of microbicide

safety are not yet available, the murine HSV-2 infection modelrepresents one of the best available models for evaluatingmicrobicide safety using susceptibility to infection as an end-

Figure 2. Schematic summary of the key events of endothelial cellactivation and leukocyte mucosalinfiltration in response to com-pound-mediated cervicovaginal ep-ithelial irritation. The inflammatory

cascade starts with release of IL-1and other cytokines by damaged or stressed epithelial cells, followed byNF-B and AP-1 mediated induc-tion of endothelial vascular adhe-sion molecules and leukocyte traf-

ficking and activation at the site of injury. The transendothelial migra-tion involves a complex sequenceof molecular and cellular events.The adhesion molecule E-selectinis involved in the early stages of leukocyte tethering and attach-ment to the endothelium, while

the vascular adhesion molecule VCAM-1 promotes late eventssuch as firm adhesion and vascular diapedesis. The transmigrating leu-kocytes move up the chemokinegradient generated by the activatedepithelium releasing more cyto-kines and shedding soluble(s) E-se-lectins, VCAM-1, CD14, and/or MPO into the vaginal secretions.Reproduced with permission fromTrifonova RT, Bajpai M, PAsicznykJM, Chandra N, Doncel GF, Fi-chorova RN. Biomarkers, 2007: 1 to

15.

Cummins JR et al.




point.55 In recent studies, candidate compounds have beenevaluated for their ability to induce inflammatory mediatorstogether with increased susceptibility to HSV infection.61,62

CHALLENGES TO IMPLEMENTING BIOMARKERSOF INFLAMMATION IN CLINICAL TRIALS

For the reasons described above, a growing number of candidate biomarkers is anticipated to arise from the researchbeing undertaken today. As these candidates will ultimatelyneed clinical validation, the collection of additional specimensin current and future clinical trials seems like a logical step.However, both biomedical and logistical factors present chal-lenges that could impact collection of specimens and interpre-

tation of assays. Biomedical factors include those variables thataffect either the detection or concentration of an analyte. Lo-gistical factors include the complexity related to collection of additional samples in large studies.

A number of biomedical factors can impact the levels of inflammatory markers in genital secretions: (1) the presence of concurrent STIs, (2) changes in microflora, (3) fluctuations of themenstrual cycle, (4) hormonal changes, and (5) the impact of coitus with or without semen deposition. For example, IL-1levels are increased in women with bacterial vaginosis, preg-nant women with normal flora, and in normal women onhormonal contraceptives with normal flora (Table 2).60,63 Theinnate factor SLPI is also increased in pregnant women or afterintercourse with semen deposition, yet is decreased in women

with bacterial vaginosis or those in menopause.6,64 These resultsfurther highlight the gaps in our knowledge of biomedical factorsused as biomarkers since (1) the impact of hormonal status onmost biomarkers is largely unknown, (2) the effect of STIs andmicroflora is incompletely understood, and (3) the impact of intercourse and semen deposition has not been clearly established.

A number of factors contribute to the complexity of adding collection of biomarkers to large scale trials. Largetrials are generally designed to be simple about specimencollection because of site burden and trial expense. Thus,collection of additional samples for ancillary studies orexploratory aims is difficult to justify. For a representativelarge study (3000 women) with 2 years follow-up, theaddition of a genital sample collected at the quarterly visits

is estimated to add 50,000 extra specimens or 170,000extra specimens for monthly visits. The cost of lab suppliesfor collection and storage of the additional monthly samples

is estimated to be as much as $500,000. This does notinclude any additional costs for extra pelvic exams, time perpatient visit, or clinical staff. In addition, extra samplingcould impact patient acceptability.

In spite of these challenges, most experts agree thatcollection of biologically relevant samples from clinical studies

for the prospective and retrospective validation of biomarkers,models, and assays is crucial to finding predictive markers of microbicide safety.

CONCLUSIONS AND FUTURE PROSPECTS

Although a number of biomarkers have been linked tomicrobicide-induced cervicovaginal inflammation, most of themare easily influenced by other physiological and pathologic pro-cesses, concurrently occurring in the genital mucosa. While manyof these markers have been measured in preclinical and clinicalassays, there are currently no data on the correlation between anyone marker and susceptibility to HIV-1 infection in humans.Although cytotoxic compounds such as N-9 and BZK can induce

reproducible biomarker profiles in preclinical assays (cell-basedand explant-based assays) and animal models (RVI and mouseHSV models), which consistently reflect a mucosal immunoin-flammatory reaction, it is not clear how consistent and predictivethese profiles may be after treatment with other microbicide can-didates. This review has focused mostly on discovery and earlydevelopment of nonclinical biomarkers of microbicide mucosalsafety. To date, the microbicides field does not have any validatedbiomarker, a gap in the knowledge base that certainly hinders therational and expeditious selection of candidates entering clinicaltrials. Furthermore, clinical biomarkers of cervicovaginal mucosalsafety, with their intrinsic higher variability, have proven to beeven more elusive.

Future directions include (1) identification of new, highlyreproducible, predictive biomarkers (nonclinical and clinical), andmodels of microbicide-induced changes in the mucosa, (2) estab-lishment of their predictive value in clinical trials, (3) assessmentof the impact of reproductive hormones, microflora, and seminalplasma on microbicide-mucosa interactions, (4) determination of microbicide effects on mucosal inflammation and innate immu-nity, and (5) study of the impact of chronic (repeated, frequent)mucosal exposure to microbicide candidates.

Given the challenges associated with implementing suchbiomarkers in large clinical trials, it is clear that only the mostrobust, validated biomarkers should be considered for use inthese trials. Therefore, a process of sequential discovery, char-acterization, and validation of new biomarkers of microbicidesafety should be implemented both at preclinical and earlyclinical stages. This process will benefit from the reevaluationof microbicide candidates tested in phase I-III clinical trials andthe investigation of available biologic samples from these trials.New, validated and highly predictive biomarkers, endpointsand models of microbicide cervicovaginal safety are essentialto the rational selection of candidates for efficacy trials.

REFERENCES1. Givan AL, White HD, Stern JE, et al. Flow cytometric analysis of

leukocytes in the human female reproductive tract: Comparisonof fallopian tube, uterus, cervix, and vagina. Am J Reprod Im-munol 1997; 38:350–359.

2. Yeaman GR, Guyre PM, Fanger MW, et al. Unique CD8 Tcell-rich lymphoid aggregates in human uterine endometrium.J Leukoc Biol 1997; 61:427–435.

3. Schaefer TM, Fahey JV, Wright JA, et al. Innate immunity in thehuman female reproductive tract: Antiviral response of uterine

TABLE 2. Conditions That Influence Levels of IL-1 and SLPIin Cervicovaginal Lavages

IL-1 SLPI

Increased in women with

Nugent score 7

Increased in pregnant

womenIncreased in pregnant womenvs. nonpregnant womenwith normal flora

Decreased in womenwith BV ortrichomoniasis

Increased among normalwomen on hormonalcontraceptives with normalflora

Increased followingintercourse withsemen deposition

Increased in the presence of STIs

Decreased withmenopause

BV indicates bacterial vaginosis.Used with permission from Sharon Hillier.





epithelial cells to the TLR3 agonist poly(I:C). J Immunol 2005;174:992–1002.

4. Schaefer TM, Wright JA, Pioli PA, et al. IL-1beta-mediated proin-flammatory responses are inhibited by estradiol via down-regulationof IL-1 receptor type I in uterine epithelial cells. J Immunol 2005;175:6509–6516.

5. Pioli PA, Weaver LK, Schaefer TM, et al. Lipopolysaccharide-

induced IL-1 beta production by human uterine macrophagesup-regulates uterine epithelial cell expression of human beta-defensin 2. J Immunol 2006; 176:6647–6655.

6. Fahey JV, Wira CR. Effect of menstrual status on antibacterialactivity and secretory leukocyte protease inhibitor production byhuman uterine epithelial cells in culture. J Infect Dis 2002;185:1606–1613.

7. Trifonova RT, Doncel GF, Fichorova RN. Polyanionic microbi-cides modify TLR-mediated innate immune responses. Antimi-crob Agents Chemother. Epub January 12, 2009.

8. Klotman ME, Rapista A, Teleshova N, et al. Neisseria gonorrhoeae-induced human defensins 5 and 6 increase HIV infectivity: Role inenhanced transmission. J Immunol 2008; 180:6176–6185.

9. Kuhn R, Lohler J, Rennick D, et al. Interleukin-10-deficient micedevelop chronic enterocolitis. Cell 1993; 75:263–274.

10. Sadlack B, Merz H, Schorle H, et al. Ulcerative colitis-likedisease in mice with a disrupted interleukin-2 gene. Cell 1993;75:253–261.

11. Sellon RK, Tonkonogy S, Schultz M, et al. Resident enteric bacteriaare necessary for development of spontaneous colitis and immunesystem activation in interleukin-10-deficient mice. Infect Immun1998; 66:5224–5231.

12. Asseman C, Mauze S, Leach MW, et al. An essential role forinterleukin 10 in the function of regulatory T cells that inhibitintestinal inflammation. J Exp Med 1999; 190:995–1004.

13. Kullberg MC, Jankovic D, Feng CG, et al. IL-23 plays a key rolein Helicobacter hepaticus-induced T cell-dependent colitis. J ExpMed 2006; 203:2485–2494.

14. Erdman SE, Rao VP, Poutahidis T, et al. CD4()CD25()regulatory lymphocytes require interleukin 10 to interrupt coloncarcinogenesis in mice. Cancer Res 2003; 63:6042–6050.

15. Rakoff-Nahoum S, Hao L, Medzhitov R. Role of toll-like recep-tors in spontaneous commensal-dependent colitis. Immunity2006; 25:319–329.

16. Wang C, Tang J, Crowley-Nowick PA, et al. Interleukin (IL)-2and IL-12 responses to Chlamydia trachomatis infection in ado-lescents. Clin Exp Immunol 2005; 142:548–554.

17. Fichorova RN, Cronin AO, Lien E, et al. Response to Neisseria

gonorrhoeae by cervicovaginal epithelial cells occurs in the ab-sence of toll-like receptor 4-mediated signaling. J Immunol 2002;168:2424–2432.

18. Herbst-Kralovetz MM, Quayle AJ, Ficarra M, et al. Quantifica-tion and comparison of toll-like receptor expression and respon-siveness in primary and immortalized human female lower genitaltract epithelia. Am J Reprod Immunol 2008; 59:212–224.

19. Yao XD, Fernandez S, Kelly MM, et al. Expression of toll-like

receptors in murine vaginal epithelium is affected by theestrous cycle and stromal cells. J Reprod Immunol 2007;75:106 –119.

20. Gill N, Davies EJ, Ashkar AA. The role of toll-like receptorligands/agonists in protection against genital HSV-2 infection.Am J Reprod Immunol 2008; 59:35–43.

21. Mayaud P, McCormick D. Interventions against sexually trans-mitted infections (STI) to prevent HIV infection. Br Med Bull2001; 58:129–153.

22. Freeman EE, Weiss HA, Glynn JR, et al. Herpes simplex virus 2infection increases HIV acquisition in men and women: System-atic review and meta-analysis of longitudinal studies. AIDS 2006;20:73–83.

23. Moodley P, Connolly C, Sturm AW. Interrelationships amonghuman immunodeficiency virus type 1 infection, bacterial vagi-nosis, Trichomoniasis, and the presence of yeasts. J Infect Dis2002; 185:69–73.

24. Uma S, Balakrishnan P, Murugavel KG, et al. Bacterial vaginosis infemale sex workers in Chennai, India. Sex Health 2005; 2:261–262.

25. Sewankambo N, Gray RH, Wawer MJ, et al. HIV-1 infectionassociated with abnormal vaginal flora morphology and bacterialvaginosis. Lancet 1997; 350:546–550.

26. van de Wijgert JH, Mason PR, Gwanzura L, et al. Intravaginalpractices, vaginal flora disturbances, and acquisition of sexuallytransmitted diseases in Zimbabwean women. J Infect Dis 2000;181:587–594.

27. Galvin SR, Cohen MS. The role of sexually transmitted diseasesin HIV transmission. Nat Rev Microbiol 2004; 2:33– 42.

28. Doncel GF, Chandra N, Fichorova RN. Preclinical assessment of the proinflammatory potential of microbicide candidates. J AcquirImmune Defic Syndr 2004; 37(suppl 3):S174–S180.

29. Beer BE, Doncel GF, Krebs FC, et al. In vitro preclinicaltesting of nonoxynol-9 as potential anti-human immunodefi-ciency virus microbicide: A retrospective analysis of resultsfrom five laboratories. Antimicrob Agents Chemother 2006;50:713–723.

30. Fichorova RN, Richardson-Harman N, Alfano M, et al. Biologi-cal and technical variables affecting immunoassay recovery of cytokines from human serum and simulated vaginal fluid: Amulticenter study. Anal Chem 2008; 80:4741–4751.

31. Nuttall JP. The use of oligonucleotide based microarray data from

human cervical and vaginal epithelial cell lines as a model forcharacterizing the potential local toxicity of candidate microbicides.In: Microbicides-2008 Conference; 2008; New Delhi, India.

32. Mesquita PM. Disruption of the epithelial barrier by cellulosesulfate: Development of a model to assess microbicide safety. In:Microbicides-2008 Conference; 2008; New Delhi, India.

33. Abner SR, Guenthner PC, Guarner J, et al. A human colorectalexplant culture to evaluate topical microbicides for the preventionof HIV infection 1. J Infect Dis 2005; 192:1545–1556.

34. Cummins JE Jr, Guarner J, Flowers L, et al. Preclinical testing of candidate topical microbicides for Anti-HIV-1 activity and tissuetoxicity in a human cervical explant culture. Antimicrob AgentsChemother 2007; 51:1770–1779.

35. Fletcher P, Kiselyeva Y, Wallace G, et al. The nonnucleosidereverse transcriptase inhibitor UC-781 inhibits human immuno-

deficiency virus type 1 infection of human cervical tissue anddissemination by migratory cells. J Virol 2005; 79:11179–11186.

36. Fletcher PS, Elliott J, Grivel JC, et al. Ex vivo culture of humancolorectal tissue for the evaluation of candidate microbicides. AIDS2006; 20:1237–1245.

37. Grivel JC, Elliott J, Lisco A, et al. HIV-1 pathogenesis differs inrectosigmoid and tonsillar tissues infected ex vivo with CCR5-and CXCR4-tropic HIV-1. AIDS 2007; 21:1263–1272.

38. Gupta P, Ratner D, Patterson BK, et al. Use of frozen-thawedcervical tissues in the organ culture system to measure anti-HIVactivities of candidate microbicides. AIDS Res Hum Retroviruses2006; 22:419– 424.

39. Shattock R, Fischetti L, Fletcher P, et al. Development of ahuman penile tissue explant model to evaluate safety and efficacyof microbicide candidates. In: 3rd IAS Conference on HIV Patho-

genesis and Treatment; July 24–27, 2005; Rio de Janeiro, Brazil.40. Ayehunie S, Cannon C, Lamore S, et al. Organotypic humanvaginal-ectocervical tissue model for irritation studies of spermi-cides, microbicides, and feminine-care products. Toxicol In Vitro2006; 20:689– 698.

41. Trifonova RT, Pasicznyk JM, Fichorova RN. Biocompatibility of solid-dosage forms of anti-human immunodeficiency virus type 1microbicides with the human cervicovaginal mucosa modeled exvivo. Antimicrob Agents Chemother 2006; 50:4005– 4010.

42. Lard-Whiteford SL, Matecka D, O’Rear JJ, et al. Recommendationsfor the nonclinical development of topical microbicides for preven-tion of HIV transmission: An update. J Acquir Immune Defic Syndr2004; 36:541–552.

43. Fichorova RN, Tucker LD, Anderson DJ. The molecular basis of nonoxynol-9-induced vaginal inflammation and its possible rele-vance to human immunodeficiency virus type 1 transmission. J In-fect Dis 2001; 184:418–428.

44. Hanson J, Posner S, Hassig S, et al. Assessment of sexuallytransmitted diseases as risk factors for HIV seroconversion in a

Cummins JR et al.




New Orleans sexually transmitted disease clinic, 1990–1998.Ann Epidemiol 2005; 15:13–20.

45. Fichorova RN. Guiding the vaginal microbicide trials with bi-omarkers of inflammation. J Acquir Immune Defic Syndr 2004;37(suppl 3):S184–S193.

46. Fichorova RN, Bajpai M, Chandra N, et al. Interleukin (IL)-1,IL-6, and IL-8 predict mucosal toxicity of vaginal microbicidal

contraceptives Biol Reprod 2004; 71:761–769.47. Milstone DS, Fukumura D, Padgett RC, et al. Mice lacking

E-selectin show normal numbers of rolling leukocytes but re-duced leukocyte stable arrest on cytokine-activated microvascularendothelium. Microcirculation 1998; 5:153–171.

48. Harashima S, Horiuchi T, Hatta N, et al. Outside-to-inside signalthrough the membrane TNF-alpha induces E-selectin (CD62E)expression on activated human CD4 T cells. J Immunol 2001;166:130–136.

49. Schenkel AR, Mamdouh Z, Muller WA. Locomotion of mono-cytes on endothelium is a critical step during extravasation. NatImmunol 2004; 5:393–400.

50. Lawn SD, Subbarao S, Wright TC Jr, et al. Correlation betweenhuman immunodeficiency virus type 1 RNA levels in the femalegenital tract and immune activation associated with ulceration of

the cervix. J Infect Dis 2000; 181:1950–1956.51. Trifonova RT, Bajpai M, Pasicznyk JM, et al. Biomarkers of leukocyte traffic and activation in the vaginal mucosa. Biomark-ers 2007; 12:608–622.

52. Bourne N, Zaneveld LJ, Ward JA, et al. Poly(sodium 4-styrenesulfonate): Evaluation of a topical microbicide gel against herpessimplex virus type 2 and Chlamydia trachomatis infections inmice. Clin Microbiol Infect 2003; 9:816 – 822.

53. Cone RA, Hoen T, Wong X, et al. Vaginal microbicides: Detect-ing toxicities in vivo that paradoxically increase pathogen trans-mission. BMC Infect Dis 2006; 6:90.

54. Tuyama AC, Cheshenko N, Carlucci MJ, et al. ACIDFORMinactivates herpes simplex virus and prevents genital herpes in amouse model: Optimal candidate for microbicide combinations.J Infect Dis 2006; 194:795–803.

55. Galen BT, Martin AP, Hazrati E, et al. A comprehensive murinemodel to evaluate topical vaginal microbicides: Mucosal inflam-mation and susceptibility to genital herpes as surrogate markers of safety. J Infect Dis 2007; 195:1332–1339.

56. John M, Keller MJ, Fam EH, et al. Cervicovaginal secretionscontribute to innate resistance to herpes simplex virus infection.J Infect Dis 2005; 192:1731–1740.

57. Venkataraman N, Cole AL, Svoboda P, et al. Cationic polypep-tides are required for anti-HIV-1 activity of human vaginal fluid.J Immunol 2005; 175:7560–7567.

58. Hazrati E, Galen B, Lu W, et al. Human alpha- and beta-defensinsblock multiple steps in herpes simplex virus infection. J Immunol2006; 177:8658– 8666.

59. Keller MJ, Guzman E, Hazrati E, et al. PRO 2000 elicits a declinein genital tract immune mediators without compromising intrinsicantimicrobial activity. AIDS 2007; 21:467–476.

60. Spandorfer SD, Neuer A, Giraldo PC, et al. Relationship of abnormal vaginal flora, proinflammatory cytokines and idiopathicinfertility in women undergoing IVF. J Reprod Med 2001; 46:806–810.

61. Mesquita PM, Wilson SS, Manlow P, et al. Candidate microbi-cide PPCM blocks human immunodeficiency virus type 1 infec-

tion in cell and tissue cultures and prevents genital herpes in amurine model. J Virol 2008; 82:6576–6584.62. Patel S, Hazrati E, Cheshenko N, et al. Seminal plasma reduces

the effectiveness of topical polyanionic microbicides. J Infect Dis2007; 196:1394–1402.

63. Sturm-Ramirez K, Gaye-Diallo A, Eisen G, et al. High levels of tumor necrosis factor-alpha and interleukin-1beta in bacterialvaginosis may increase susceptibility to human immunodefi-ciency virus. J Infect Dis 2000; 182:467–473.

64. Novak RM, Donoval BA, Graham PJ, et al. Cervicovaginal levelsof lactoferrin, secretory leukocyte protease inhibitor, and RAN-TES and the effects of coexisting vaginoses in human immuno-deficiency virus (HIV)-seronegative women with a high risk of heterosexual acquisition of HIV infection. Clin Vaccine Immunol2007; 14:1102–1107.



bio markers of cervicovaginal inflammation for the.4[1]

Documents