Microbicides are products intended to be applied to the vaginal or rectal mucosa to prevent the sexual transmission of HIV-1 and potentially other sexually transmitted pathogens. One of the greatest challenges in assessing the suitability of an active agent for use as a microbicide is the lack of surrogate markers of efficacy. In HIV-1 treatment, one of the early achievements was change in patient viral load proved to be a surrogate for the patient's clinical response to antiretroviral treatment. However, in HIV-1 prevention research, there have been no recognized surrogates of efficacy, which are relevant for clinical trial outcomes. Preclinical laboratory assays and animal models used to evaluate products for activity against HIV-1 have been poor predictors of effectiveness in HIV prevention trials. Both ectocervical and colorectal tissues obtained from surgical specimens have been used to evaluate the impact of drugs in tissues for preclinical drug testing.1–4 The tissue is exposed to drug and HIV-1 in the laboratory to follow HIV-1 infection, or lack thereof, as an indicator of treatment efficacy. These explant models rely on human tissue, which retain the architecture and relevant immune cells and thus provide a model that closely resembles intact female genital tract and colonic tissues.
The explant model was further developed by Anton et al5,6 during their phase 1 studies where vaginal microbicides were applied rectally before sampling colonic tissue by endoscopy. After product use, the biopsy tissue was then challenged with HIV-1 in the laboratory. This model was termed “ex vivo challenge.” The ex vivo challenge model is considered to be one of the closest surrogates to in vivo infection. Therefore, it is a potential bioindicator of microbicide product effectiveness. The participants in these studies used UC781-containing, tenofovir-containing, or placebo gels daily for up to 1 week. At the end of the week, colonic biopsies were taken and the ones designated for ex vivo challenge were brought to the laboratory and exposed to HIV-1. Colonic tissue taken from the 2.5% UC781 and 1% tenofovir gel participants showed a significant reduction in HIV-1 p24 release compared with the colonic tissue taken from the placebo gel participants. These encouraging results with colonic tissues are tempered by the high levels of inter- and intrasubject variability in HIV-1 replication.7 Increasing the reliability of the ex vivo challenge, model will provide better predictions of clinical trial outcomes. To do this, evaluating the effects of assay methods and biopsy collection techniques on variability in tissue infection is needed. First, the capacity of HIV-1 to replicate in ectocervical vs. vaginal tissue is unknown. Because the vagina has a greater surface area than the ectocervix, some investigators recommended using only vaginal tissue biopsies, whereas others suggested that the greater numbers of CD4+ immune cells in the ectocervix would make it a better choice. Second, the impact of female genital tract pre-procedure preparation that includes biopsy site disinfection and the application of topical anesthetic needs to be evaluated. It is unknown if either procedure would have an impact on HIV-1 infection of tissue. The effects of biopsy location (vaginal or ectocervical) and pre-procedure topical anesthesia (lidocaine) and disinfectant (chlorhexidine) on ex vivo biopsy infection were tested on both weight-adjusted and non–weight-adjusted viral replication as measured by p24 in culture supernatant and viral infection as measured by p24 immunohistochemistry (IHC). The goal of the present study was to define the optimal tissue and collection procedures for use of female genital tract tissue in the ex vivo challenge assay for future incorporation into early clinical trials.
Unless otherwise indicated, tissue culture base medium was purchased from Mediatech Inc. (Manassas, VA); supplements were purchased from Gemini Bio-products (West Sacramento, CA) or Lonza (Walkersville, MD); and interleukin-2 was purchased from Roche (Indianapolis, IN). The optimal challenge dose of HIV-1 may vary with virus strain and the laboratory growing the virus and determining the viral titer. The challenge virus should be empirically determined before using in a clinical study. For the current work, concentrated HIV-1BaL was purchased from Advanced Biotechnologies Inc. (Columbia, MD). Dilutions of HIV-1BaL were made in RPMI-1640 and stored at −80°C. Tissue culture infectious dose 50% (TCID50) was determined by the Reed and Muench method for phytohemagglutinin-activated peripheral blood mononuclear cells.8
Mucosal Tissue Collection
Paired ectocervical and vaginal tissue biopsies were collected from healthy women who consented to the procedure (University of Pittsburgh IRB PRO10110377). Forty-two women had paired ectocervical and vaginal tissue biopsies collected. Of these, a total of 28 women, 14 women per group, had 2 additional biopsies from each site collected without 5% lidocaine or chlorhexidine treatment. For women who underwent 4 biopsies with and without a topical preparation, after insertion of the speculum, attention was turned to the right vaginal wall. A Tischler biopsy instrument was positioned over a fold of vaginal tissue in the proximal vagina and a full-thickness biopsy was collected. The Tischler instrument was then positioned at the external opening of the cervix with 1 arm resting in the opening of the cervix, and a biopsy was obtained. The topical preparation was then applied to the left vaginal wall and cervical face, and the second set of vaginal and cervical biopsies were obtained. Silver nitrate was applied to the vaginal biopsy sites and Monsels was applied to the cervical biopsy sites for hemostatis. The tissue was placed in a conical tube with transport medium (L-15 medium supplemented with 10% fetal bovine serum, 100 µg/mL streptomycin, 100 U/mL penicillin, 100 mM L-glutamine, and 2.5 µg/mL amphotericin B), placed on wet ice, and brought to the laboratory within 30 minutes of collection.
Ex Vivo Challenge Assay
Tissues were placed into individual wells of a 48-well plate with 1 mL of complete medium (Dulbecco's Modified Eagle Medium supplemented with 10% human AB serum, 100 µg/mL streptomycin, 100 U/mL penicillin, 100 mM L-glutamine, 1% nonessential amino acids, and 100 U/mL interleukin-2) with 5 × 104 TCID50 HIV-1BaL and cultured in a humidified incubator at 37°C, 5% CO2. After 2 hours, the tissue was washed by passing through 4 changes of Hank's Balanced Salt Solution without calcium/magnesium and then placed in a new well of a 48-well plate with 1 mL of fresh complete medium without HIV-1. Every 3–4 days of culture, 700 µL of medium was collected and stored at −80°C, and fresh complete medium was replenished. Cultures were maintained for 21 days based on our work with ectocervical explants.2 However, maintaining cultures through earlier time points, such as 14 days of culture, can also be done.4 After 21 days of culture in this study, the tissue was weighed and placed in formalin to process for HIV-1–infected cells expressing p24 by IHC. The culture supernatant was evaluated for soluble HIV-1 p24 by enzyme-linked immunosorbent assay (Perkin–Elmer Life Sciences Inc., Boston, MA).
Weight-adjusted (pg/mg) and non–weight-adjusted (pg/mL) p24 measurements were summed across test days 4 through 21 to provide cumulative p24 pg/mg and cumulative p24 pg/mL end points. The soft end point was calculated for weight-adjusted and non–weight-adjusted p24 end points,9 and its calculation template is available at http://www.alphastatconsult.com/id3.html. Cumulative and soft p24 end points (pg/mL and pg/mg) were log10 transformed before analyses. IHC end points were provided for 36 of 42 ectocervical and 40 of 42 vaginal biopsy tissues. The weight was compared between paired ectocervical and vaginal tissues using the Wilcoxon matched-pairs signed-rank test.
A randomized block design analysis of variance (PROC MIXED) was used to compare the log-transformed virus growth measurements for paired ectocervical vs. vaginal biopsy challenge results, and at each biopsy location, both with and without either lidocaine or chlorhexidine. Demographic and behavioral measures were included as covariates, that is, binary behavioral measurements of age, smoking, hormonal oral contraceptive pill use, condom use, intrauterine device use, and sexual activity were provided. IHC rates were compared by a 2 × 2 Fisher exact text between ectocervical vs vaginal biopsies, and for each biopsy type, ± lidocaine and ± chlorhexidine. Intercorrelations between biopsy end points were performed using Pearson and logistical analyses. An alpha value of 0.05 and SAS (version 9.3; Cary, NC) were used for all analyses.
Study Population and Tolerability of the Biopsy Procedures
The population characteristics of the 42 volunteers who participated in the study and provided tissues are shown in Table 1. The study participants were predominately white, and the median age was 29 years (range, 21–45 years). Fewer than 10% were current smokers, and the majority of women were using some form of birth control. None of the women used Depo-Provera or had a tubal ligation. Women reported that both vaginal and ectocervical biopsies were well tolerated. For the 14 study participants who had 4 biopsy samples obtained with and without lidocaine, no difference in perception between the sampling procedures was noted. Thus, there was no difference in the patient acceptability of biopsy with or without anesthetic.
HIV-1 Infection of Ectocervical and Vaginal Tissues
To address whether ectocervical and vaginal tissues are infected and have a similar capacity to replicate HIV-1, paired tissues from 42 participants were obtained. One participant's tissue became contaminated during the culture period; therefore, paired tissues from 41 participants were evaluated. Because of variable tissue sizes obtained by biopsy, the p24 results were analyzed without and with correction of tissue weight at study end. The end of the 21-day culture period was chosen because mucus is often attached to the tissue at time of collection, and additional manipulation was not desired at the beginning of the study. The median tissue weights were significantly different (P < 0.001) between ectocervical (median, 8.68 mg; range, 0–36.59 mg) and vaginal (13.29; 0–123.56). Overall, replication of HIV-1 in ectocervical tissue was significantly higher (P < 0.005) than in vaginal tissue as measured by p24 release into the culture supernatant (Fig. 1A; Table 2). Weight-adjusted and non–weight-adjusted p24 end points were highly correlated (P < 0.0001) for the cumulative (Fig. 2A) and soft (Fig. 2B) end point p24 analyses. However, the weight-adjusted and non–weight-adjusted end points showed more variability at the lower p24 levels for vaginal tissue as evidenced by the increased scatter of the red data points. None of the covariate factors were found to influence HIV-1 replication for either tissue type.
Detection of p24 expressing cells by IHC confirmed HIV-1 infection of the tissue. Ectocervical and vaginal tissues from 36 and 40 donors, respectively, were available at study end for IHC. No significant differences were found in IHC-positive tissues; 31 of 36 (86%) ectocervical p24-positive tissues (Fig. 1B) compared with 36 of 40 (90%) vaginal p24-positive (Fig. 1C) tissues (P = 0.7279).
Effect of Chlorhexidine and Lidocaine on HIV-1 Infection
To determine whether application of disinfectant or anesthetic modified HIV-1 infection in tissue explants, a subset of women underwent additional procedures in addition to the initial 2 biopsies. Additional procedures included either cleaning the tissue with chlorhexidine before obtaining 2 additional biopsies (14 women) or numbing the tissue with 5% lidocaine gel before obtaining 2 additional biopsies (14 women). There was no significant effect of either procedure on HIV-1 p24 production in ectocervical tissue (Table 2). Although lidocaine had no effect on HIV-1 p24 production in vaginal tissue, chlorhexidine-treated tissue showed a modest increase in HIV-1 p24 production when adjusted for tissue weight (P < 0.05; Table 2) but no effect on HIV-1 p24 production when not adjusted for tissue weight.
p24 End Point Correlations With IHC
To define a point estimate for HIV-1 infection of the tissue, both p24 end points were correlated with IHC p24-postive tissues. There were significant positive relationships between ectocervical tissue and all p24 end point measurements [Figs. 3A, B; soft end point (P < 0.05) not shown]. Vaginal tissue significantly correlated with the weight-adjusted p24 end point measurements [Fig. 3C; soft end point (P = 0.0359) not shown] but not with the non–weight-adjusted p24 end point measurements [Fig. 3D; soft end point (P = not significant) not shown]. To estimate the p24 end points that provide a 95% probability of an IHC p24-positive tissue, logistic regression point estimates were interpolated. The p24 logistic regression point estimates for ectocervical tissue were determined to be 3.43 and 2.98 log10 (pg/mg of tissue) for cumulative p24 and soft p24 end points, respectively. Ectocervical p24 levels were approximately 1 log10 higher than vaginal tissue, which were determined to be 2.5 and 1.96 log10 (pg/mg of tissue) for cumulative p24 and soft p24 end points, respectively. Overall, the cumulative p24 was approximately 0.5 log10 higher than the soft p24 end points in both tissues.
This study has demonstrated that freshly obtained tissue biopsy samples from the female genital tract were consistently infected with HIV-1 and so this model can be used as tissue explants to assess susceptibility to HIV-1 when challenged “ex vivo.” The ex vivo challenge assay was used to test vaginal topical microbicide candidates after rectal application.5,6 The initial study collected colonic tissue located 10 cm and 30 cm from the anal verge to define HIV-1 infection and replication which were similar at both anatomic sites.6 The translation of this assay to female genital tract tissue required that vaginal and cervical tissue be evaluated for their capacity to be infected and replicate HIV-1. Although both ectocervical and vaginal tissues were infected with HIV-1 at a similar frequency, ectocervical tissue replicated HIV-1 to approximately 1 log10 higher levels. Compared with the colonic tissue, female genital tract tissue responds to HIV-1 infection differently as seen by vaginal tissue having modest p24 levels. Moreover, pre-procedure preparation of the 2 mucosal tracts is quite different. Before collection of colonic tissue, the participant is asked to fast the night before, and then, the next morning, the lumen is rinsed with an isotonic enema to remove any solid material that may remain. Before collection of female genital tract tissue, women commonly have the vaginal vault “washed” with an antiseptic solution and may have the biopsy site numbed with a topical anesthetic. Topical antiseptic solutions often contain chlorhexidine, which is a potent antibacterial compound and is known to inactivate HIV-1 and herpes simplex virus.10 Because the immune and nervous systems communicate with each other, anesthetics are known to dampen the immune response.11 Collectively, these procedures could impact the ex vivo challenge assay by interfering with HIV-1 infection of the tissue. Chlorhexidine and lidocaine pretreatments showed no overall significant impact on HIV-1 infection or replication in ectocervical or vaginal tissue. However, chlorhexidine pretreatment had a modest increase in HIV-1 replication in vaginal tissue when adjusted for tissue weight. As the collection of tissue without these extraneous procedures was tolerated by the 42 women in this study, it is suggested that chlorhexidine and lidocaine treatments are not performed before tissue collection for the ex vivo challenge assay.
Despite using the same sized Tischler forceps to collect the ectocervical and vaginal tissues, there was significant variability in the size of the biopsies likely due to varying thickness of the vaginal folds. Because mucus is often collected along with the biopsy and additional manipulations through washing were not desired, biopsies were weighed at study end (day 21) of culture when mucus would have dissipated. Ectocervical tissue had a significantly lower weight than vaginal tissue. Three biopsies, 2 ectocervical and 1 vaginal, had no discernible tissue at study end. Regardless of the variability in the tissue weight, the cumulative and soft p24s were highly correlated for non–weight-adjusted and weight-adjusted end points with the ectocervical tissue producing approximately 1 log10 more HIV-1 p24 than the vaginal tissue. Although immune cells populate the lower genital tract, the cervix has more immune cells, specifically CD4+ T cells, than the vagina, and the cell populations do not seem to be influenced through the menstrual cycle.12,13 Although no systematic study has been done to compare immune cell activation between the ectocervix and vagina, the CD4+ T-cell subpopulations in the cervix expressing α4β7 (mucosal homing receptor) also express higher levels of CCR5 and CD69.14 Although we were not able to define the immune cell populations and their activation status in this study due to the limited number of biopsies collected, the published data would support our results showing similar infection rates between ectocervical and vaginal tissues but greater HIV-1 replication occurring in the ectocervical tissue due to more activated immune targets. Adjusting for the weight of the tissue may be beneficial, especially for vaginal tissue, as the p24 end point correlations to p24-positive tissue significantly correlated when weight adjusted. To note, the 2 studies5,6 using colonic tissue for the ex vivo challenge assay did not weight adjust their p24 results. The colon has the largest numbers of immune cells in the body, and most of those cells are activated15 so weight adjustment may not be as critical for interpretation, similar to the ectocervical results reported here. A limitation of the work presented was not collecting the tissue weight before placing it in culture. Although our intent was to eliminate washing steps, which would exclude mucus from the tissue, pre- and postculture weighing may be warranted to accurately define the impact tissue weight has on HIV-1 replication. Nonetheless, these results show the utility of both ectocervical and vaginal tissue types for the ex vivo challenge assay.
The use of female genital tract tissues presents additional challenges compared with colonic tissues due to the number of female genital tract tissue biopsies that can feasibly be collected. Because the colon has a large surface area, taking up to 20 biopsies is typical, whereas the numbers of biopsy samples which can be obtained at a single time point for female genital tract samples is 4: 2 vaginal and 2 ectocervical. This limits the within-subject sampling across the genital tract for any clinical trial evaluating an HIV-1 preventative. Tissue is usually required for immune cell distribution and numbers, biomarkers such as gene arrays, drug levels (pharmacokinetics), in addition to drug protection (pharmacodynamics) so prioritization of tissue usage should be made early in the study to maximize tissue use. Additionally, healing of the biopsy area may take up to 10 days, so frequent tissue collection is limited and restricted for ethical reasons. A recent study demonstrated that taking multiple colonic biopsies did not increase uptake of radioactive small molecules—similar in size to HIV-1 particles.16 This type of study should be conducted for the female genital tract after collection of vaginal/ectocervical biopsies to ensure that epithelial trauma does not influence the normal homeostasis of the genital tract. As discussed above, there are differences in immune cell distribution through the genital tract. Although the cell numbers do not seem to vary over the menstrual cycle, exogenous hormones from contraceptive use has been shown to increase activated CD4+/CCR5+ cells in the vagina.17 In nonhuman primates, the use of progestin was shown to reduce the vaginal epithelial thickness and increase Simian Immunodeficiency Virus susceptibility.18 Indeed, depot medroxyprogesterone acetate injections in female macaques are done to aid in the transmission of SIV in experiments testing the effectiveness of vaccines and microbicides. Although there has been some indication that the use of hormonal contraception may increase a women's risk of HIV-1 infection,19 the use of hormonal contraception by women in this study did not influence the infection or replication of HIV-1. However, further work is needed to systematically define the role of hormonal contraceptives in HIV-1 infection of mucosal tissue. Although additional preparatory procedures of the vaginal vault such as the use of chlorhexidine and lidocaine had no impact on HIV-1 infection, chlorhexidine modestly increased HIV-1 replication in vaginal tissue when adjusted for weight but not in ectocervical tissue. Chlorhexidine is a potent antibacterial compound with documented antiviral activity10 and is a common ingredient in mouthwashes and topically applied products. Chlorhexidine was shown to enhance HSV transmission and induce prolonged susceptibility to chlamydia in mice.20,21 These additional preparatory procedures should be eliminated to allow for better interpretation of the ex vivo challenge assay results from clinical studies using topical microbicide products.
Although this bioindicator model has shown promise evaluating some topical formulations rectally, the incorporation of the ex vivo challenge assay may not be useful in all early clinical trials. Proximity of the clinical site location to the laboratory performing the ex vivo challenge assay and the activity of the drug candidate being tested also must be considered. Currently, only fresh tissue has been used for this assay. The use of frozen tissue for HIV-1 infection has not yielded reproducible results. Cervical tissue seems to remain intact,22 but the capacity of HIV-1 to replicate in fresh compared with frozen tissue was not tested. Another study showed that colorectal tissue quickly degrades after freeze/thawing procedures.23 These studies suggest that the freezing and thawing process of tissue needs to be optimized and standardized before using frozen tissue for the ex vivo challenge assay. Moreover, the osmotic changes that occur to the cells in the tissue during freezing and thawing may compromise the drug levels in the tissue by inadvertently flushing the cells of drug.24 Therefore, the drug candidate being tested also needs to be considered before using the ex vivo challenge assay. For example, luminally active drugs such as PRO 2000, Griffithsin, or monoclonal antibodies would not be expected to be effective at preventing HIV-1 infection of the tissue because the drug presumably would be removed during specimen collection and transport before HIV-1 exposure. Non-nucleoside reverse transcriptase inhibitors, which have no cellular transporters and thus freely diffuse in and out of the cell, could be partially or completely removed as well. Consequently, the use of the ex vivo challenge assay is limited to laboratories in close proximity to the clinical trial site and potentially to drugs that are active within the cell.
Fresh ectocervical and vaginal tissues from young healthy women to be used for HIV-1 infection can be incorporated into early clinical trials to evaluate candidate drugs. Ectocervical tissue replicates HIV-1 approximatey 1 log10 higher than vaginal tissue, despite HIV-1 infecting the tissues equally. Although the inclusion of preparatory procedures for tissue collection such as cleansing and numbing the genital tract vault had no to modest impact on HIV-1 infection and replication, the participants did tolerate the tissue collection without them. Thus, these preparatory procedures are not necessary and not including them would potentially reduce some of the variability that is inherent in the assay. We are now incorporating the ex vivo challenge assay into several clinical trials as a bioindicator of the drug activity in the tissue after in vivo use.
The authors would like to thank the volunteers for their willingness to participate in research.
1. Abner SR, Guenthner PC, Guarner J, et al.. A human colorectal explant culture to evaluate topical microbicides
for the prevention of HIV infection. J Infect Dis. 2005;192:1545–1556.
2. Cummins JE Jr, Guarner J, Flowers L, et al.. Preclinical testing of candidate topical microbicides
for anti-human immunodeficiency virus type 1 activity and tissue toxicity in a human cervical explant culture. Antimicrob Agents Chemother. 2007;51:1770–1779.
3. Fletcher PS, Elliott J, Grivel JC, et al.. Ex vivo culture of human colorectal tissue for the evaluation of candidate microbicides
. AIDS. 2006;20:1237–1245.
4. Greenhead P, Hayes P, Watts PS, et al.. Parameters of human immunodeficiency virus infection of human cervical tissue and inhibition by vaginal virucides. J Virol. 2000;74:5577–5586.
5. Anton PA, Cranston RD, Kashuba A, et al.. RMP-02/MTN-006: a phase 1 rectal safety, acceptability, pharmacokinetic and pharmacodynamic study of tenofovir 1% gel compared to oral tenofovir disoproxil fumarate. AIDS Res Hum Retroviruses. 2012;28:1412–1421.
6. Anton PA, Saunders T, Elliott J, et al.. First phase 1 double-blind, placebo-controlled, randomized rectal microbicide trial using UC781 gel with a novel index of ex vivo efficacy. PLoS One. 2011;6:e23243.
7. Richardson-Harman N, Mauck C, McGowan I, et al.. Dose-response relationship between tissue concentrations of UC781 and explant infectibility with HIV type 1 in the RMP-01 rectal safety study. AIDS Res Hum Retroviruses. 2012;28:1422–1433.
8. Reed LJ, Muench H. A simple method of estimating fifty per cent endpoints. Am J Hyg. 1938;27:493–497.
9. Richardson-Harman N, Lackman-Smith C, Fletcher PS, et al.. Multisite comparison of anti-human immunodeficiency virus microbicide activity in explant assays using a novel endpoint analysis. J Clin Microbiol. 2009;47:3530–3539.
10. Baqui AA, Kelley JI, Jabra-Rizk MA, et al.. In vitro effect of oral antiseptics on human immunodeficiency virus-1 and herpes simplex virus type 1. J Clin Periodontol. 2001;28:610–616.
11. Yu HP. Role of anesthetic agents on cardiac and immune systems. Shock. 2011;36:532–541.
12. Pudney J, Quayle AJ, Anderson DJ. Immunological microenvironments in the human vagina and cervix: mediators of cellular immunity are concentrated in the cervical transformation zone. Biol Reprod. 2005;73:1253–1263.
13. Ildgruben AK, Sjoberg IM, Hammarstrom ML. Influence of hormonal contraceptives on the immune cells and thickness of human vaginal epithelium. Obstet Gynecol. 2003;102:571–582.
14. McKinnon LR, Nyanga B, Chege D, et al.. Characterization of a human cervical CD4+ T cell subset coexpressing multiple markers of HIV susceptibility. J Immunol. 2011;187:6032–6042.
15. Anton PA, Elliott J, Poles MA, et al.. Enhanced levels of functional HIV-1 co-receptors on human mucosal T cells demonstrated using intestinal biopsy tissue. AIDS. 2000;14:1761–1765.
16. Fuchs EJ, Grohskopf LA, Lee LA, et al.. Detecting Rectal Epithelial Disruption Using Radioisotopes: A Simple Test to Identify Potential HIV Microbicide Toxicity. Microbicide 2008. New Delhi, India; 2008. Abstract #209.
17. Chandra N, Thurman AR, Anderson S, et al.. Depot medroxyprogesterone acetate increases immune cell numbers and activation markers in human vaginal mucosal tissues. AIDS Res Hum Retroviruses. 2013;29:592–601.
18. Marx PA, Spira AI, Gettie A, et al.. Progesterone implants enhance SIV vaginal transmission and early virus load. Nat Med. 1996;2:1084–1089.
19. Heffron R, Donnell D, Rees H, et al.. Use of hormonal contraceptives and risk of HIV-1 transmission: a prospective cohort study. Lancet Infect Dis. 2012;12:19–26.
20. Cone RA, Hoen T, Wong X, et al.. Vaginal microbicides
: detecting toxicities in vivo that paradoxically increase pathogen transmission. BMC Infect Dis. 2006;6:90.
21. Achilles SL, Shete PB, Whaley KJ, et al.. Microbicide efficacy and toxicity tests in a mouse model for vaginal transmission of Chlamydia trachomatis. Sex Transm Dis. 2002;29:655–664.
22. Gupta P, Ratner D, Patterson BK, et al.. Use of frozen-thawed cervical tissues in the organ culture system to measure anti-HIV activities of candidate microbicides
. AIDS Res Hum Retroviruses. 2006;22:419–424.
23. McGowan I, Tanner K, Elliott J, et al.. Nonreproducibility of "snap-frozen" rectal biopsies for later use in ex vivo explant infectibility studies. AIDS Res Hum Retroviruses. 2012;28:1509–1512.
24. Mazur P. Freezing of living cells: mechanisms and implications. Am J Physiol. 1984;247:C125–C142.