Pharmacokinetics and Pharmacodynamics of Tenofovir Reduced-Glycerin 1% Gel in the Rectal and Vaginal Compartments in Women: A Cross-Compartmental Study with Directly Observed Dosing

Abstract

Background

Evidence is lacking regarding whether vaginal pre-exposure prophylaxis with topical tenofovir (TFV) reduces the risk of rectal HIV acquisition.

Setting

Bronx, NY

Methods

MTN-014 was a phase 1 cross-over, randomized sequence trial comparing the cross-compartment pharmacokinetics and pharmacodynamics of daily TFV reduced-glycerin (RG) 1% gel following 14 days each of rectal and vaginal application, with directly observed dosing and a 6-week washout period between phases.

Results

Fourteen HIV-uninfected women enrolled; 91% of doses were observed and 13 women completed all study procedures. TFV and tenofovir diphosphate (TFV-DP) were detected in most samples collected from the dosing compartment. After vaginal dosing, TFV was detected in 10/14 samples of rectal fluid (RF) (median 4.4 ng/sponge) and 1/13 rectal tissue samples (0.2 ng/mg); TFV-DP was detected in 2/13 rectal tissue samples at 59.8 and 76.5 fmol/mg. After rectal dosing, TFV was detected in 9/14 samples of vaginal fluid (median 1.1 ng/swab) and in 6/14 vaginal tissue samples (median below limit of quantification); TFV-DP was detected in 3/14 vaginal tissue samples at 17.3, 87.6, 77.1 fmol/mg. Neither cervicovaginal lavage fluid nor RF collected 24 hours after rectal or vaginal dosing resulted in a statistically significant suppression of viral replication.

Conclusion

In this study of 14 days each of vaginal and rectal application of TFV RG 1% gel, we found only a small degree of cross-compartment distribution of TFV in RF and vaginal fluids and no pharmacodynamic activity in ex vivo testing. While high TFV concentrations in the dosing compartment may be protective, low cross-compartment tissue concentrations are not likely to be protective.

INTRODUCTION

To provide broadly effective HIV prevention for women, prevention approaches that protect against both vaginal and rectal HIV acquisition are required. Oral pre-exposure prophylaxis (PrEP) with tenofovir disoproxil fumarate-emtricitabine (TDF-FTC) appears to meet this need, as it reduces the risk of both vaginal and rectal HIV acquisition,^1–3 although daily adherence may be necessary to achieve vaginal protection.^4,5 While the evidence is inconsistent, topical PrEP with tenofovir (TFV) 1% gel appears to reduce the risk of vaginal acquisition, especially after adjusting for poor adherence in post hoc analyses.^6–8 Evidence is lacking, however, regarding whether vaginal PrEP with TFV reduces the risk of rectal HIV acquisition. Given the highly variable rates of heterosexual anal intercourse reported among women worldwide, ^9,10 much of which is unprotected, ^11–14 it is critical to understand the cross-compartment distribution of topical vaginal and rectal TFV.

Non-human primate studies of TFV gel applied vaginally and rectally have shown evidence of cross-compartment distribution of TFV between the vagina and rectum, in each case down a 1-2 log₁₀ gradient.¹⁵ The concentrations achieved through cross-compartment distribution were similar to concentrations shown to provide protection to non-human primates challenged with SHIV vaginally ^16–19 or rectally.²⁰ In one clinical trial, TFV gel applied vaginally resulted in quantifiable rectal fluid concentrations, with higher concentrations than those seen after oral dosing.²¹ While these initial findings suggest that vaginal dosing might provide some rectal mucosa protection in the setting of anal intercourse, more directly supporting data, including cross-compartment tenofovir diphosphate (TFV-DP) tissue concentrations, are needed. Consequently, a Phase 1, cross-over, randomized trial in women was conducted to describe the steady-state, cross-compartmental pharmacokinetics (PK) and pharmacodynamics (PD) of the investigational study product, TFV reduced glycerin (RG) 1% gel, applied vaginally or rectally. The trial also assessed the safety of vaginally applied TFV RG 1% gel. This formulation is better tolerated than the high glycerin (HG) formulation when applied rectally,²² however, its safety when applied vaginally had not been evaluated.

METHODS

Study Design

MTN-014 was a phase 1, randomized, open-label, vaginal and rectal gel dosing cross-over study of TFV RG gel. Endpoints included drug concentrations in fluid and tissue from the vagina and rectum, safety of TFV RG gel, and acceptability of daily directly observed dosing (DOD). The study was conducted at the Bronx Prevention Center Clinical Research Site (Bronx, NY) under a protocol approved by the Columbia University Institutional Review Board, and in accordance with the Helsinki Declaration of 1975, as revised in 2000.

The investigational TFV RG gel study product was supplied by CONRAD (Arlington, VA) and manufactured by DPT Laboratories (San Antonio, TX). The gel pH was 4.0-5.5, the osmolality was approximately 836 mmol/kg and it was supplied in 4 mL pre-filled, single dose applicators.

Women were eligible if they were between 21 and 45 years of age (inclusive), HIV-uninfected, and willing, during the study period, to use effective contraception and abstain from using non-study vaginal or rectal products. Exclusion criteria included hepatitis B virus surface antigen seropositivity, active sexually transmitted infection, use of pre- or post-exposure HIV prophylaxis, pregnancy or breastfeeding, and significant gynecological, blood chemistry or hematologic abnormalities. At screening, vaginal fluid (VF) and rectal fluid (RF) were collected for nucleic acid amplification testing (NAAT) for Neisseria gonorrhoeae and Chlamydia trachomatis and VF was collected for a rapid trichomonas test. Blood was collected for safety laboratory evaluations (complete blood count, creatinine, alanine aminotransferase and aspartate aminotransferase) and syphilis serology. Urine was collected for pregnancy testing.

Participants were randomized 1:1 to one of two sequences of gel administration periods (vaginal-rectal or rectal-vaginal) (Figure 1). Each period was 14 days long, separated by a 6-week washout period that enabled sample collection at a similar point in the participants’ menstrual cycles. Participants presented daily for a DOD visit during each product-use period, including weekends, so that study staff could directly observe study product insertion. Each participant received two applicators per 14-day product-use period for use at home in case a DOD visit was not possible.

Figure 1.

Study procedures during period 1 and period 2 of MTN 014

At the end of each product-use period, approximately 24 hours (h) after the final (14^th) dose, blood, VF, RF, vaginal and rectal tissue biopsies, and endocervical cells (ECC) were collected. At the beginning of the second product-use period, VF and RF were collected. Safety was assessed at each follow-up visit by physical, pelvic and rectal examinations. Participant attitudes about DOD and daily clinic visits were assessed via face-to-face interview at enrollment and at the end of each product-use period.

Specimen Collection Methods

Blood was collected for plasma and peripheral blood mononuclear cell (PBMC) isolation. VF was collected by Dacron swab (posterior fornix) and cervicovaginal lavage (CVL) was collected after administering 10 mL of normal saline. ECC were collected by QIAGEN Digene cytobrush (Gaithersburg, MD) and, after elution from the cytobrush, total cell count (including squamous cells) and percent of viable cells were determined using trypan blue and a hemacytometer. RF was collected with a Merocel (Mystic, CT) sponge inserted anoscopically. Two 2–4 mm vaginal biopsies were collected (Tischler Rotating Biopsy Punch, Gynex Corporation, #1200 or 1008), one from each lateral vaginal wall. Four rectal biopsies were collected 15 cm from the anal verge via sigmoidoscope (Wilson-Cook Max forceps, #G54167, cup diameter 3.3 mm).

Laboratory Methods

TFV and TFV-DP concentrations were determined by previously described liquid chromatographic-tandem mass spectrometric methods validated²³ for each biological specimen source and relevant fluid collection device.^24,25 TFV and TFV-DP concentrations were determined from a single biopsy from each source, vaginal or rectal, to minimize inter-biopsy variation. For TFV, lower limits of quantification (LLOQ) for each matrix were: plasma, 0.31 ng/mL; CVL, 5 ng/mL; VF, 0.625 ng/swab; RF, 1.25 ng/sponge; tissue, 0.05 ng/sample. TFV-DP was quantified in PBMC, ECC, and tissue homogenates using a previously described indirect assay which measured TFV following isolation of TFV-DP and enzymatic dephosphorylation to TFV.²⁵ The LLOQ of TFV-DP for all matrices was 50 fmol/sample. TFV-DP results were normalized to number of cells analyzed for PBMC and ECC and reported as fmol/million cells. Based on the number of cells analyzed, the median TFV-DP LLOQs for PBMC and ECC were 24 and 12 fmol/million cells, respectively. Tissue TFV and TFV-DP values were normalized to biopsy weight and reported as ng/mg and fmol/mg, respectively. Tissue-specific LLOQs were calculated by dividing the LLOQ, by the weight of each analysed tissue sample homogenate, and then calculating the median value. CVL concentrations were not adjusted for dilutional volume.

CVL and RF were tested for anti-HIV activity using the TZM-bl assay.^22,26 Cellular debris was removed from the CVL by centrifugation and the supernatant was aliquoted and stored at ≤−70°C until tested. Sponge eluates were obtained by placing the sponge in the upper portion of a SpinX column (Corning Inc., Corning, NY). After centrifugation, eluates were aliquoted and stored at ≤−70°C until tested. Two dilutions of each CVL and rectal sponge eluate were tested in triplicate along with HIV-1_BaL. Modest dilutions of mucosal fluids (1:4 for CVL and 1:37.5 for RF) were made to optimize the pre-dose innate activity against the presumptive post-dose drug-associated activity. Saline was used as a control for the mucosal fluid. Anti-HIV activity was measured by luminescent assay (BrightGlo; Promega Corp., Madison, WI). Inhibition was determined based on deviations from the HIV-1-only controls and presented as the percent inhibition.

Statistical Analysis

Descriptive statistics were used to summarize study population demographics, PK, and PD outcomes. Concentrations of TFV and TFV-DP in the dosing compartment (e.g., vaginal concentration after vaginal dosing) were summarized and compared with concentrations in the cross-compartment (e.g., vaginal concentration after rectal dosing). A linear mixed model was used to test for interaction between product and use period. For PD data, percentage of change from pre-dose values to post-dose values was computed for CVL and RF, and compared between dosing compartment and cross-compartment delivery. In all randomized participants, number and percentage of adverse events (AE) were tabulated by severity, category, and relationship to study product.

RESULTS

Study Population and Visit Performance

Fourteen women, recruited primarily from family planning and gynecology clinics in Bronx, New York, enrolled in the study between May and November 2014. The participants’ mean age was 33.7 years; most were unmarried (79%), college-educated (57%), identified their race as Black or African American (50%) and their ethnicity as Hispanic (50%), and earned their own income (71%). Participants completed all study procedures except for one participant who opted not to undergo biopsies at the end of her second (vaginal) product-use period. As a result, there were 13, rather than 14, tissue samples collected after the 14 day vaginal product-use periods.

Study staff directly observed 358 of 392 expected doses (91%), including 100% of the final doses. A total of 32 doses were reported as used at home and 2 doses were reported as missed. The 32 non-observed doses occurred at a variety of times across the two product-use periods, without evidence of clustering at the beginning or end (data not shown).

Pharmacokinetics

Plasma TFV concentrations were quantifiable in 10/14 samples (71%) 24 h after the end of vaginal and rectal product-use periods, and median TFV plasma concentrations were similar, 0.58 ng/mL and 0.82 ng/mL, respectively (Table 1). Only one PBMC sample, collected after rectal dosing, had quantifiable TFV-DP. Although no participants had quantifiable TFV in plasma or RF samples at the end of the washout phase, VF samples from two participants starting rectal product-use periods had quantifiable TFV (327.5 and 0.938 ng/swab).

Table 1.

Pharmacokinetics of Sampled Compartments by Dosing Route

Dosing Route	Compartment	Analyte (units)	LLOQ	#Quantifiable Samples	Concentration, median (IQR1)1
Vaginal Application	Plasma	TVF† (ng/mL)	0.31	10/14	0.58(BLQ§, 1.3)
	PBMCs	TFV-DP‡ (fmol/106 cells)	50 fmol/sample	0/14	BLQ (BLQ, BLQ)
	Vaginal fluid	TVF(ng/swab)	0.625	14/14	7138 (1370, 24375)
	CVL	TVF (ng/mL)	5	13/14	4485 (2410, 48900)
	Vaginal tissue2,3	TVF (ng/mg)	0.0046	12/13	8.5 (1.0, 44.8)
		TFV-DP (fmol/mg)	4.61	12/13	166 (37, 2377)
	Endocervical cells	TFV-DP (fmol/106 cells)	50 fmol/sample	8/14	126.3 (BLQ, 421.7)
	Rectal fluid	TFV (ng/sponge)	1.25 ng/sponge	10/14	4.4 (BLQ, 7.1)
	Rectal tissue2,4	TVF (ng/mg)	0.0270	1/13	0.2
		TFV-DP (fmol/mg)	27.03	2/13	59.8; 76.5
Rectal Application	Plasma	TVF (ng/mL)	0.31	10/14	0.82 (BLQ, 1.2)
	PBMCs	TFV-DP (fmol/106 cells)	50 fmol/sample	1/14	48.32
	Vaginal fluid	TVF (ng/swab)	0.625	9/14	1.1 (BLQ, 2.8)
	CVL	TVF (ng/mL)	5	3/14	9.2; 8.5; 43
	Vaginal tissue2,3	TVF (ng/mg)	0.0046	6/14	BLQ (BLQ, 0.16)
		TFV-DP (fmol/mg)	4.61	3/14	17.3; 87.6; 77.1
	Endocervical cells	TFV-DP (fmol/106 cells)	50 fmol/sample	0/14	BLQ (BLQ, BLQ)
	Rectal fluid	TVF (ng/sponge)	1.25 ng/sponge	12/14	202 (55, 2205)
	Rectal tissue2,4	TVF (ng/mg)	0.0270	12/14	3.0 (0.7, 10.9)
		TFV-DP (fmol/mg)	27.03	10/14	196 (BLQ, 550)

Notes:

^*Inter-Quartile Range;

^†Tenofovir;

^§Below the Limit of Quantitation;

^‡Tenofovir Diphosphate

¹Values represent median of all collected samples followed, in parentheses, by the interquartile range (IQR). When three or fewer samples were quantifiable, all values are shown without parentheses.

²The overall LLOQ of TFV was 0.05 ng/mg tissue and the overall LLOQ of TFV-DP was 50 fmol/mg of tissue. The tissue-specific LLOQ of TFV and of TFV-DP were calculated by dividing the overall LLOQ by the median weight of the biopsy samples from each anatomic location. As the median tissue weights obtained after each dosing route were similar, the LLOQs for vaginal and rectal tissue are based on the average of the tissue weights obtained after each dosing route.

³Mean duration of time between the final dose at the end of each product-use period and vaginal tissue collection was 21.6 h (SD 2.6 h).

⁴Mean duration of time between the final dose at the end of each product-use period and rectal tissue collection was 20.4 h (SD 2.7 h)

Median tissue-specific LLOQs of TFV and TFV-DP were each about six times smaller in vaginal tissue as compared to rectal tissue (Table 1), reflecting the larger median weights of the vaginal versus rectal tissue samples (10.85 mg versus 1.85 mg, data not shown).

After vaginal dosing, TFV was detected in all (14/14) VF and nearly all (13/14) CVL samples, with median concentrations of 7138 ng/swab and 4485 ng/mL, respectively (Table 1). These cannot be directly compared since the swab was not corrected for sample weight and the CVL was not corrected for dilutional effects of the lavage. TFV and TFV-DP were detected in nearly all, 12/13 (92%), vaginal tissue samples, with median concentrations of 8.5 ng/mg and 166 fmol/mg, respectively, and TFV-DP was detected in more than half, 8/14 (57%), of ECC, with a median concentration of 126.3 fmol/10⁶ ECC.

Detection of vaginally applied TFV in the rectal compartment varied according to sample type. While TFV was quantifiable in most (10/14) RF samples with a median concentration of 4.4 ng/sponge, TFV and TVF-DP were detected in only 1/13 (0.2 ng/mg) and 2/13 (59.8 and 76.5 fmol/mg) rectal tissue samples, respectively. The TFV rectal-vaginal concentration ratio in the one sample that had quantifiable rectal tissue TFV was 0.2 to 8.5 ng/mg or 1:42; the TFV-DP rectal:vaginal concentration ratios in the two samples that had quantifiable rectal tissue TFV-DP were 59.8 to 166 fmol/mg and 76.5 to 166 fmol/mg or approximately 1:3 and 1:2, respectively. As these rectal:vaginal ratios were calculable for only 3/26 samples, these estimates should be viewed with caution and may not represent the unmeasurable ratios in other participants.

After rectal dosing, similar to vaginal dosing, TFV was detected in most of dosing compartment samples (Table 1). TFV was detected in 12/14 RF samples, with a median concentration, of 202 ng/sponge, and in 12/14 rectal tissues, with a median concentration of 3.0 ng/mg. TFV-DP was detected in 10/14 (71%) rectal tissues, with a median concentration of 196 fmol/mg.

As with vaginal dosing, detection of rectally applied TFV in the cervicovaginal compartment varied according to sample type. While TFV was detected in most (9/14) VF samples, with a median concentration of 1.1 ng/swab, TFV was detected in only 3/14 CVL samples. In vaginal tissue homogenates, drug was detected in fewer than half, TFV in 6/14 and TFV-DP in 3/14 samples, with concentration medians largely BLQ; and in none of the ECC samples. Median LLOQs for TFV and TFV-DP in vaginal tissue samples were 0.0046 ng/mg and 4.61 fmol/mg, respectively; and despite a median cytobrush count of 5.0×10⁶ cells, the median TFV-DP in ECC samples was BLQ. The vaginal:rectal tissue ratios for TFV after rectal dosing for the 6 samples that had quantifiable TFV in vaginal tissue was at least 0.0046 (the LLOQ) to 3.0 ng/mg or 1:652; and the vaginal:rectal tissue ratio for TFV-DP for the 3 samples that had quantifiable vaginal TFV-DP was at least 17.3 to196 fmol/mg or 1:11. As above, these vaginal:rectal ratios should be viewed with caution as the ratios were calculable for only 9/28 samples.

Except for rectal sponge concentrations (p = 0.043), there were no statistically significant interactions between route of dosing and period of use in this trial. As above, due to the limited number of vaginal tissue samples with quantifiable concentrations, paired analyses of each participant’s findings were not performed.

HIV Inhibition

Baseline CVL samples demonstrated modest anti-HIV activity (41% ± 14%) (Table 2), reflecting natural innate anti-viral activity. After 14 days of vaginal product use, the anti-HIV activity increased to 71% (SD 29%), but was not significantly different (Table 2). No significant change in the inhibition of HIV was observed with RF collected after vaginal dosing or in CVL or RF after rectal dosing.

Table 2.

Pharmacodynamics: % inhibition of HIV, Mean (Standard Deviation)

	Vaginal use phase			Rectal use phase
	Pre- dose	Post- dose	% change, post- versus pre- dosing	Pre- dose	Post- dose	% change,
						post- versus pre-dosing
Cervical vaginal lavage fluid, 1:4 dilution	41 (14)	71 (29)	42%	39 (16)	42 (18)	7%
Rectal fluid (sponge), 1:37.5 dilution	51 (39)	59 (29)	14%	50 (38)	43 (50)	−14%

Safety

One participant experienced weight loss during the rectal phase. Initially categorized as a Grade 2 AE, it progressed to a Grade 3 event, resulting in a product hold of two days; it was due to social stress and not related to product use. Two other Grade 2 AEs were observed during the rectal phase, bacterial vaginosis (Grade 2) and vaginal trichomoniasis (Grade 2), each in different participants, and neither event was deemed related to product use.

Acceptability of DOD

Most participants (23/28) reported always inserting study product herself, rather than having study staff insert it, and most reported feeling “comfortable” (12/28) or “very comfortable” (7/28) with clinicians observing them. Most reported that they liked (7/28) or were neutral (16/28) about the required daily study site visits.

DISCUSSION

Women may be exposed to HIV through both vaginal and rectal intercourse, yet few studies have assessed the bidirectional pharmacokinetics of antiretroviral ^15,21 or other pharmacologic agents between vaginal and rectal compartments. Understanding cross-compartment PK and PD of vaginally and rectally applied HIV prevention products is crucial to determine their potential to provide HIV protection, regardless of the compartment to which product is applied or route of sexual exposure.⁶ In this open-label, cross-over, randomized Phase 1 trial, we found that when used for 14 days either vaginally or rectally, TFV RG gel was safe and acceptable in adult women. DOD was acceptable in the context of this clinical trial.

We applied a comprehensive anatomic sampling strategy to understand the distribution and time course of TFV and TFV-DP in cervicovaginal and rectal fluids and tissue and to enable multi-compartment modeling in secondary analyses. Approximately 24 h after the end of each 14-day dosing period, cross-compartment detection of TFV was more evident in fluid samples than tissue samples, due largely to increased LLOQ in tissue samples, especially for TFV-DP. In addition, cross-compartment tissue TFV (but not TFV-DP) was detected more frequently after rectal dosing than after vaginal dosing, based on the proportion of cross-compartment tissue samples with quantifiable concentrations. On the other hand, ratios of cross-compartment tissue concentrations for both TFV and TFV-DP were larger after vaginal dosing compared to rectal dosing.

This is the first study to assess vaginal use of the RG formulation of TFV gel. The RG and HG formulations of TFV gel have very similar physical and chemical properties except for glycerin content (5% versus 20%) and osmolality (836 versus 3111 mmol/kg).²² Prior clinical studies evaluating vaginal use of TFV gel used the HG formulation,^8,27,28 however, the RG formulation offers important advantages over the HG formulation as epithelial integrity is better preserved while drug release and anti HIV-1 activity are comparable.²² A comparison of findings after vaginal dosing of TFV RG gel in this study with findings after vaginal dosing of TFV HG gel in MTN-001, which had similar anatomic and temporal sampling, indicates largely overlapping concentration distributions for serum/plasma (Table 3). Vaginal tissue TFV, vaginal tissue TFV-DP, and CVL TFV concentrations after vaginal dosing appeared higher after TFV HG gel in MTN-001 when compared to TFV RG gel in MTN-014. This might be accounted for by the earlier sampling time after TFV HG gel versus TFV RG gel dosing (~13 h versus ~24 h), but this should only affect CVL and tissue TFV modestly and not affect TFV-DP due to the much longer TFV-DP half-life.²⁹ When comparing daily rectal TFV RG gel dosing between MTN-014 and the one-week dosing arm of CHARM-01³⁰, which both used similarly timed post-dose sampling, a mixed picture emerges: serum/plasma, RF TFV and CVL TFV concentrations were higher in CHARM-01; rectal tissue TFV concentrations were similar; and rectal tissue TFV-DP concentrations were lower in CHARM-01, with 25% of samples BLQ in both studies and a wider range of values in MTN-014 (Table 3). For the vaginal compartment, too few samples had quantifiable drug after rectal dosing to make a meaningful comparison.

Table 3.

Comparison of MTN-014 PK results with vaginal and rectal PK results from other selected studies of TFV

Study	N	Route Freq. Duration	Serum TFV C24h ng/mL	Vaginal Tissue TFV ng/mg	Vaginal Tissue TFV-DP fmol/mg	Rectal Tissue TFV ng/mg	Rectal Tissue TFV-DP fmol/mg	Rectal fluid TFV ng/sponge	CVL Fluid TFV ng/mL
MTN-014 (RG, vaginal dosing)	14	qd × 14d DOD	0.58 [BLQ, 1.3]	8.5 [1.0, 44.8]	166 [37, 2377]	0.2; (1/13)	59.8, 76.5; (2/13)	4.4 (BLQ, 7.1]	4485 [2410, 48900]
MTN-014 (RG, rectal dosing)	14	qd × 14 d DOD	0.82 [BLQ, 1.2]	BLQ [BLQ, 0.16]	17.3, 87.6, 77.1; (3/14)	3.0 [0.7, 10.9]	196 [BLQ, 550]	202 [55, 2205]	9.2, 8.5, 43; (3/14)
MTN-001*21 (HG, vaginal dosing)	144	qd × 42 d	0.67 [BLQ, 2.09]	113 [27, 265]	1807 [591, 5860]	–	–	119 (53–2150); (12/12)	15.5 × 104 [3.0 × 104, 40.5 × 104]
MTN-001* (TDF, oral dosing)	144	qd × 42 d	65 [14, 103]	0.15 [BLQ, 0.27]	BLQ [BLQ, BLQ]	–	–	20 (7, 404); (11/12)	269 [BLQ, 10078]
HPTN-066**37 (TDF, oral dosing)	13*/49	qd × 35 d DOD	52 [49, 56];(13/13)	0.02, 0.18; (2/2)	BLQ, 41 (2/2)	5.42 [0.21, 20.54]	206 [BLQ,595]	1617; (1/4)	625, 6290; (2/2)
CHARM 0130 (RG, rectal dosing)	12	R, qd × 7 d	6.0 (4.3–7.1) plasma Cmax	-	-	1.4 [0.7, 3.7]	5.2 [BLQ, 12.8]	9.4 × 105 ng/mL [4.3 × 105–1.4 × 106)	1,824; 2,460

Abbreviations: C₂₄: concentration 24 h after dose; DOD: directly observed dosing; RG: reduced glycerin gel; HG: high glycerin gel; qd: daily dosing; BLQ: below level of quantification;

PK values are median [IQR]. When three or fewer quantifiable samples are available, all values are shown separated by comma and followed by semi-colon, then followed by the number of quantifiable samples/total samples in parentheses.

^*In MTN-001, values are all pre-dose, estimated to be ~13 h after the prior dose (unpublished data). N=144 for serum and CVL, 72 for vaginal tissue, and 12 for RF; CVL values in MTN-001 are reported originally as corrected for dilution, 20-fold, but reported uncorrected here to allow comparison to MTN-014.

^**HPTN 066 had 13 participants in the daily dosing arm which included serum samples, and 2 men and 2 women in the tissue and luminal fluid sampling arms.

Oral dosing of TDF protects individuals at risk of HIV acquisition via either vaginal or rectal exposure to HIV, presumably because vaginal and rectal concentrations of TFV-DP are sufficient to prevent HIV infection,^1,2 though it must be noted that vaginal and rectal TFV-DP concentrations have not definitively been shown to be the relevant site and analyte for HIV protection. However, the estimated tissue TFV-DP concentrations among the primary PrEP randomized clinical trials, including vaginal dosing in the VOICE gel arm, offer a good explanation of trial outcomes and we would therefore rank tissue TFV-DP as the most influential determinant of PrEP outcomes, regardless of dosing route.⁵ If one compares vaginal tissue concentrations of TFV and TFV-DP associated with daily rectal dosing of TFV RG gel in this study with oral TDF dosing arms in MTN-001, an oral regimen proven protective of predominantly vaginal HIV exposure risk in several RCTs, most samples after both dosing routes are BLQ, rendering a comparison of “protective” concentrations impossible. A comparison of rectal tissue TFV and TFV-DP concentrations 24 h after daily vaginal dosing in this study with concentrations achieved 24 h after daily oral dosing in HPTN 066, a regimen protective for rectal HIV exposure risk, reveals the daily oral regimen results in higher concentrations. Vaginal dosing of TFV RG gel may therefore be less protective against rectal transmission of HIV than oral dosing of TDF. We believe the pre-dose 24 h TFV-DP tissue concentrations to be most relevant to a daily dosing scenario because this represents the time of lowest concentration.

We used the TZM-bl cell assay^22,26,31 to assess the relationship between TFV concentration and HIV viral replication in fluid samples taken from both vaginal and rectal compartments. Neither CVL nor RF collected after rectal or vaginal dosing resulted in statistically significant suppression of viral replication. Based on in vitro testing of HIV-1_BaL, the 50% effective concentration (EC₅₀) of TFV in the TZM-bl assay is 611 ng/mL. While some CVL samples collected after vaginal dosing had TFV concentrations above this, the modest dilution used for the lavage in the assay (1:4) likely resulted in loss of TFV-associated anti-HIV activity. Thus, while there was an increase in anti-HIV activity post-vaginal dosing, overall it was not significant. The concentration of TFV in CVL after rectal dosing and in RF after either dosing regimen failed to achieve the in vitro EC₅₀ concentration, however, lack of activity in the TZM-bl assay does not necessarily mean that vaginal or rectal use of TFV RG gel would fail to provide cross-compartment protection from HIV acquisition 24 h after application. Recent work by others indicates that lactobacillus-dominant vaginal flora may enhance ^32,33 and inflammatory cytokines may diminish ^34,35 the protective efficacy of vaginal TFV gel in healthy HIV-negative women. Such factors may be relevant to the findings in this study and additional analyses are underway to examine the relationship between vaginal flora, cytokines and TFV concentrations.

This study had several strengths: the DOD schedule maximized adherence, the cross-over design allowed each participant to serve as her own control, and sampling 24 h after 14 days of daily dosing ensured samples were collected after steady-state concentrations had been established. This sampling time point, based on evidence of persistent TFV in vaginal aspirates and tissue for up to 24h after the last of 14 daily doses in a TFV HG 1% gel study,³⁶ sheds light on the lowest (pre-dose) concentrations in daily dosing regimens. Finally, rectal biopsy samples were collected at an anatomic location associated with the highest concentration of cell-free and cell-associated HIV surrogates in coital simulation studies.²⁹

The small number of tissue samples with quantifiable levels limited our ability to perform within-participant comparisons. In addition, the rectal tissue-specific LLOQs, six times higher than the vaginal tissue-specific LLOQs, limited our ability to detect TFV and TFV-DP in rectal tissue. This difference in LLOQs may explain, at least in part, the higher proportion of rectal tissue samples with unquantifiable concentrations after vaginal dosing as compared to the proportion of vaginal tissue samples after rectal dosing. To note, we have reported VF TFV and RF TFV results in ng/swab and ng/sponge, respectively. Because these concentrations were not normalized to collected fluid weights, while general statements regarding relative amount of drug in each matrix can be made, a direct comparison between studies is challenging due to potential differences in absolute fluid volume acquired by each collection device. Finally, our inferences about cross-compartment gradients and levels of protection as assessed from published literature come from several different studies – a combination of RCTs and PK-focused studies – rather than studies that directly assess both concentration and protection in the same set of participants.

In conclusion, we found that TFV RG gel, applied rectally and vaginally for 14 days each, resulted in low cross-compartment fluid TFV concentrations. These concentrations were often below the quantitative limits of detection, making it difficult to compare the concentrations with the anatomically corresponding levels associated with daily oral dosing regimens known to be protective. While it is highly likely that, once licensed, a vaginal microbicide will also be used rectally, our data suggest that application of TFV RG 1% gel directly to the vagina is required for protection from vaginal HIV exposures and, similarly, application directly to the rectum is required for rectal HIV protection.