Progestin-Containing Contraceptives Alter Expression of Host Defense-Related Genes of the Endometrium and Cervix

Abstract

Epidemiological studies indicate that progestin-containing contraceptives increase susceptibility to HIV, although the underlying mechanisms involving the upper female reproductive tract are undefined. To determine the effects of depot medroxyprogesterone acetate (DMPA) and the levonorgestrel intrauterine system (LNG-IUS) on gene expression and physiology of human endometrial and cervical transformation zone (TZ), microarray analyses were performed on whole tissue biopsies. In endometrium, activated pathways included leukocyte chemotaxis, attachment, and inflammation in DMPA and LNG-IUS users, and individual genes included pattern recognition receptors, complement components, and other immune mediators. In cervical TZ, progestin treatment altered expression of tissue remodeling and viability but not immune function genes. Together, these results indicate that progestins influence expression of immune-related genes in endometrium relevant to local recruitment of HIV target cells with potential to increase susceptibility and underscore the importance of the upper reproductive tract when assessing the safety of contraceptive products.

Introduction

Contraceptives are utilized by more than 660 million women worldwide, with hormonal methods accounting for at least 20%.¹ In Sub-Saharan Africa, over 8 million use injectable hormonal contraceptives, mainly depot medroxyprogesterone acetate (DMPA).¹ Intrauterine devices, including the levonorgestrel intrauterine system (LNG-IUS), are popular and highly effective methods of long-acting reversible contraception, and nonhormonal IUS are used by over 160 million women worldwide.¹ Depot medroxyprogesterone acetate and the LNG-IUS have complex systemic and local mechanisms of action on the endometrium, ranging from progestational early effects to atrophy with long-term use.^2,3 However, epidemiological evidence indicates that DMPA may increase male to female HIV transmission,⁴ although effects of other long-acting, progestin-only methods have not been evaluated extensively. The mechanisms underlying these epidemiologic observations are unclear.

For male to female HIV transmission, the female reproductive tract (FRT) is a site of viral introduction, and mounting evidence suggests that the endometrium and cervical transformation zone (TZ, zone between the ecto- and endocervix) are potential portals of HIV entry.^5,6 In contrast to the lower FRT (vagina and ectocervix), which is lined by a multilayered squamous epithelium,⁷ the upper FRT (TZ, endocervix, and uterus) consists of a single-layered columnar epithelium. The latter is a major site of mucosal lymphoid tissue and is the first line of defense against pathogens. For example, the upper FRT epithelia secrete innate immune factors, including secretory leukocyte peptidase inhibitor and defensins, and inflammatory cytokines. Additionally, endometrial cells express the HIV coreceptors CXCR4 and CCR5.^5,8 In cycling women, progesterone promotes secretory transformation of the endometrium along with an influx of leukocytes, increased production of innate immune factors, and immune tolerance. In the absence of pregnancy and with declining progesterone levels, there is an influx of macrophages and the secretion of inflammatory cytokines locally that promote tissue desquamation and menses.⁹ Interestingly, the greatest risk of sexual acquisition of HIV is postulated to occur in the secretory phase of the cycle, indicating that progesterone may induce changes in the local immune milieu of the upper FRT that increase this susceptibility.^6,10

Although the lower FRT mucosa is the most intensively investigated site of HIV entry in male to female transmission, with numerous in vitro and in vivo studies,¹¹ upper FRT tissues are plausible portals of HIV entry and accessed by fluids deposited intravaginally.^5,6,11–15 Understanding upper FRT responses to exogenous progestins is important in the context of a possible role for them in susceptibility to HIV transmission and to better understand how vaccinations and other treatments may function in women. Herein, in a cross-sectional study of DMPA and the LNG-IUS users, we assessed the effects of these progestins on whole-genome transcriptomes of upper FRT tissues. We found moderate changes in cervical TZ and profound changes in the expression of immune-related genes of the endometrium, suggesting that mediators of innate immunity and populations of HIV-susceptible cells may be altered in the upper FRT in vivo by exposure to systemic and local progestins.

Methods

Study Design and Human Tissue Acquisition

This report is based on a cross-sectional study of DMPA and LNG-IUS users and women using no hormonal contraception, all of whom consented to study participation and procedures using a human subjects protection protocol that was approved by the University of California San Francisco Human Subjects Protection Committee. All biopsies were taken for the sole purpose of research, with no other medical indications. Samples were obtained during the mid-secretory phase from the nonhormonal contraception comparison group, 7 to 10 days after a positive home urine luteinizing hormone test (ClearBlue urine Ovulation Test, www.clearblue.com) and verified by a plasma progesterone level on the day of sampling (Table 1).The DMPA and LNG-IUS users were sampled after a minimum of 6 months of method use and did not have phase-based urine sampling. Participants were 18 to 45 years old, with intact uteri, negative HIV serologies and negative urine nucleic acid amplification tests for Chlamydia trachomatis and Neisseria gonorrhoeae, and were free from clinically evident vaginitis or other acute infections. Women who donated tissue for the control group did not have a history of abnormal bleeding. Rare spotting occurred in DMPA users and there was variability in spotting/bleeding among LNG/IUS users. Participants were asked to refrain from vaginal intercourse or to use unlubricated condoms for 10 days prior to tissue collection. Women with recent or unresolved cervical intraepithelial neoplasia, current use of steroidal or nonsteroidal anti-inflammatory drugs, recent pregnancy (within a year), and who were breast-feeding were excluded. All participants had a negative urine pregnancy test on the day of tissue collection. Participants from the nonhormonal contraception comparison group had a history of regular menstrual cycles (21-35 days) and at least 3 normal menstrual cycles since discontinuation of any previous hormonal or intrauterine device contraception.

Table 1.

Participant Information.

ID#	Group	Age	Cycle Day	Days Post LH Surge	Duration of Use, days	Serum P4	Endo Bx	Cervix Bx
PR-A	Control	43	18	7	N/A	9.4	Y	N
PR-B	Control	44	21	8	N/A	9.7	N	Y
PR-C	Control	28	22	6	N/A	5.9	N	Y
PR-D	Control	40	20	7	N/A	3.2	Y	Y
PR-E	Control	43	26	9	N/A	16.6	Y	Y
PR-F	Control	41	19	7	N/A	6.7	Y	Y
PR-G	Control	34	24	8	N/A	9.3	Y	N
PR-H	Control	32	20	7	N/A	9.9	Y	N
PR-I	Control	26	26	9	N/A	6	N	Y
PR-J	Control	27	18	6	N/A	6.4	Y	Y
PR-K	Control	26	24	8	N/A	7.7	N	Y
PR-L	Control	28	21	8	N/A	9.6	N	Y
PR-M	Control	43	23	9	N/A	2.5	Y	Y
PR-N	Control	34	21	9	N/A	16.9	N	Y
PR-O	Control	27	21	7	N/A	12.4	N	Y
PR-P	Control	38	31	8	N/A	7.8	Y	Y
PR-01	Control	25	25	8	N/A	10.8	Y	Y
PR-02	Control	36	26	ND	N/A	8.8	N	Y
PR-04	Control	28	24	9	N/A	7.7	N	Y
PR-08	Control	38	27	11	N/A	2.3	Y	Y
PR-09	Control	34	27	11	N/A	2.9	Y	Y
PR-13	Control	27	24	12	N/A	6.5	N	Y
PR-17	Control	43	23	12	N/A	3.9	N	Y
PR-03	DMPA	28	N/A	ND	462	N/A	N	Y
PR-14	DMPA	23	N/A	ND	776	N/A	N	Y
PR-15	DMPA	24	N/A	ND	1019	N/A	N	Y
PR-27	DMPA	37	N/A	ND	274	N/A	Y	Y
PR-29	DMPA	31	N/A	ND	1378	N/A	Y	Y
PR-30	DMPA	24	N/A	ND	861	N/A	N	Y
PR-32	DMPA	26	N/A	ND	811	N/A	Y	Y
PR-33	DMPA	32	N/A	ND	3251	N/A	N	Y
PR-34	DMPA	27	N/A	ND	431	N/A	Y	Y
PR-37	DMPA	38	N/A	ND	449	N/A	N	Y
PR-38	DMPA	23	N/A	ND	386	N/A	Y	Y
PR-40	DMPA	32	N/A	ND	369	N/A	N	Y
PR-42	DMPA	21	N/A	ND	1503	N/A	Y	Y
PR-43	DMPA	22	N/A	ND	421	N/A	Y	Y
PR-44	DMPA	23	N/A	ND	250	N/A	Y	Y
PR-05	LNG-IUS	28	N/A	ND	299	N/A	Y	Y
PR-06	LNG-IUS	24	N/A	ND	464	N/A	Y	Y
PR-07	LNG-IUS	28	N/A	ND	1019	N/A	Y	Y
PR-10	LNG-IUS	21	N/A	ND	308	N/A	N	Y
PR-11	LNG-IUS	23	N/A	ND	272	N/A	Y	Y
PR-12	LNG-IUS	25	N/A	ND	400	N/A	Y	Y
PR-18	LNG-IUS	32	N/A	ND	578	N/A	N	Y
PR-19	LNG-IUS	30	N/A	ND	1313	N/A	N	Y
PR-20	LNG-IUS	28	N/A	ND	745	0.5	Y	Y
PR-22	LNG-IUS	27	N/A	ND	803	11.9	N	Y
PR-23	LNG-IUS	26	N/A	ND	408	2.6	N	Y
PR-25	LNG-IUS	18	N/A	ND	755	3.4	Y	Y
PR-26	LNG-IUS	26	N/A	ND	231	0.5	Y	Y
PR-31	LNG-IUS	40	N/A	ND	758	0.5	Y	Y
PR-35	LNG-IUS	23	N/A	ND	443	12.1	Y	Y
PR-36	LNG-IUS	23	N/A	ND	520	<0.5	Y	Y
PR-39	LNG-IUS	44	N/A	ND	311	9.3	Y	N
PR-41	LNG-IUS	26	N/A	ND	242	0.7	Y	Y

Abbreviations: LH, luteinizing hormone; Y, yes; N, no; DMPA, depot medroxyprogesterone acetate; LNG-IUS, levonorgestrel intrauterine system; N/A, not available; ND, not determined.

Tissue Collection

Endometrial tissues were collected using Pipelle aspirators (Cooper Surgical, Trumbull, Connecticut). After 1 to 2 sampling insertions, tissue samples were expelled into Nunc cryovials (Roskilde, Denmark) and flash frozen in acetone and dry ice. Cervical TZ tissues were collected via punch method, placed in a Nunc vial, and flash frozen.

Systemic Progesterone Levels

Plasma progesterone was measured in control samples collected on the day of tissue collection by coated tube assays (Siemens Healthcare Diagnostics, Los Angeles, California) performed by the University of Virginia National Institutes of Health Specialized Cooperative Centers Program in Reproduction and Infertility Research Ligand Assay and Analysis Core (Table 1).

RNA Isolation and Complementary DNA Microarray

RNA was extracted from frozen tissues using Trizol Reagent (Life Technologies, Grand Island, New York) and purified using the NucleoSpin RNA II kit (Macherey-Nagel, Bethlehem, Pennsylvania) as described.¹⁶ Total RNA as well as protein (A260/A280) and organic compound (A260/A230) contamination were measured using the Nanodrop (Thermo Fisher Scientific Inc., Waltham, Massachusetts). Agilent 2100 Bioanalyzer (Agilent Technologies, Santa Clara, California) was used to verify that RNA quality met the minimum criteria for array-based complementary DNA (cDNA) generation. Samples were further processed for hybridization to the Affymetrix whole-genome Human Gene 1.0 ST arrays at the Gladstone Institutes Genomics Core according to the manufacturer’s protocol (Affymetrix, Inc, Santa Clara, California) as described.¹⁷

Gene Expression Data Processing and Statistical Analysis

The .cel data files were uploaded to the National Center for Biotechnology Information Gene Expression Omnibus Database (Series Accession GSE60129). For each sample, intensity values for all probe sets in the GeneChip operating software (Affymetrix) were imported into GeneSpring GX version 12.6 (Agilent Technologies) and processed using the robust multiarray analysis algorithm for background adjustment, normalization, and log2 transformation of perfect match values.^9,18 Intensity values in each condition and tissue type were normalized to the median of all samples. Gene lists were generated using two-way analysis of variance and Benjamini-Hochberg multiple testing correction for false discovery rate and filtered with a ≤1.5-fold change cutoff and P ≤ .05. Principal component analysis (PCA) and hierarchical clustering were performed as described previously.^9,18

Fluidigm-Based Microfluidic Quantitative Reverse Transcriptase Polymerase Chain Reaction Validation of DE Genes

For progestin versus control comparisons, differential expression validation by quantitative reverse transcriptase polymerase chain reaction (q-RT-PCR) was performed (TZ: 17 genes DMPA and 13 LNG-IUS; endometrium: 52 DMPA and 69 LNG-IUS). Primers (Supplemental Table 1) were purchased (Fluidigm Corp, South San Francisco, California) and cDNAs reserved from the microarray (endometrium: 7 control, 12 LNG-IUS, and 9 DMPA; cervix: 7 control, 13 LNG-IUS, and 12 DMPA) were enriched for these targets according to manufacturer’s instructions. Briefly, primers were pooled and combined with 250 ng cDNA and TaqMan PreAmp Master Mix (Applied BioSystems, Grand Island, New York), then amplified with Stratagene Mx3005P (Agilent Technologies). Products were treated with USB 10 un/uL exonuclease I (Affymetrix) and diluted 1:20. According to the manufacturer’s protocol, sample and assay mixtures were added to 48.48 (TZ) or 96.96 (endometrium) Dynamic Array IFC chips, loaded with the integrated fluidic circuits (IFC) controllers and run with the BioMark HD system (Fluidigm Corp). Data were imported to Microsoft Excel (Microsoft, Redmond, Washington), and ΔΔ CT values, average fold change, and standard deviation were calculated. Outliers were removed using Dixon Q test.¹⁹

Gene Ontology and Pathway Analyses

Gene ontology and functional annotations were evaluated using Ingenuity Pathway analysis (IPA; Ingenuity Systems, Redwood City, California), into which RefSeq IDs and fold changes of differentially expressed genes in each comparison were imported. Inhibition or activation of pathways were predicted for functional groups of genes based on collective messenger RNA expression levels, and significance was determined using the right-tailed Fisher exact test. The P values reflected the number of analysis-specific genes in a given pathway compared with the total number of occurrences of these genes in all pathways in the Ingenuity knowledge base. Results are shown for pathways with a bias-corrected Z score (reported as z=) <−1.0 or >1.0 (except for upstream regulators in cervix DMPA); however, only Z scores < −2.0 or > 2.0 (P < .05) were predicted to be inhibited or activated, respectively.

Correlation Between q-RT-PCR and Microarray Data

The correlation between microarray and q-RT-PCR differential gene expression was evaluated with nonparametric Spearman ρ and Kendall τ tests using StatView 5.0.1 (SAS Institute Inc, Cary, North Carolina). For both tests, a more positive value indicates greater agreement between microarray and q-RT-PCR differential expression values. The P values were based on a 2-tailed null hypothesis of no association.

Results

Principal Component Analysis and Hierarchical Clustering of Endometrial and Cervical Samples

The PCA revealed that samples clustered into 2 major subdivisions, endometrium and cervix (Figure 1). In general, control samples from each tissue clustered more closely with each other than did progestin-treated ones. Unsupervised hierarchical clustering analyses based on the combined gene list fold change (FC > 1.5, P < .05) derived from pairwise comparisons yielded a dendrogram of sample clustering and a heatmap of gene expression. The clustergram (Figure 2), similar to the PCA, revealed a main branching between endometrium and cervix and subbranches between controls and progestin-treated samples in the same tissue.

Figure 1.

Principal component analysis.

Figure 2.

Hierarchical clustering.

Effects of Progestins on Endometrial Gene Expression

Gene expression comparisons were made between endometrial tissue from women using 1 of the 2 target progestin contraceptives and controls during the mid-secretory (peak progesterone) phase. With DMPA versus control, somatostatin (SST), matrix metallopeptidase 7 (MMP7), and hemoglobin β (HBB) were the genes that demonstrated the greatest increase in expression, and matrix metallopeptidase 26 (MMP26), uroplakin 1B (UPK1B), and hydroxysteroid 17-β dehydrogenase 2 (HSD17B2) were the genes that showed the greatest reduction in expression. With LNG-IUS, IGFBP1, MMP7, and prolactin (PRL) demonstrated the greatest increase in expression, and homogentisate 1,2-dioxygenase (HGD), MMP26, and UPKB1 had the largest decrease in expression. High expression of IGFBP1 and PRL and LEFTY2 with both progestins confirmed the progestational effects of these drugs.^3,20 In total, there were 609 genes in LNG-IUS and 408 in DMPA which were differentially expressed when compared to tissues from unexposed women. There were 241 genes associated with treatment with either progestin, and 368 and 167 genes unique to LNG-IUS and DMPA, respectively, versus the unexposed controls (Figure 3). Full lists of differentially expressed genes are given in Supplemental Table 2A and B, and those with q-RT-PCR-validated expression levels are given in Tables 2 and 3.

Figure 3.

Venn diagrams.

Table 2.

Correlation of Array and q-RT-PCR: Endometrium DMPA Versus Control.^a,b

Gene	Array	qRT-PCR	Gene	Array	qRT-PCR
SST	11.07	82.16	AQP9	1.65	6.72
MMP7	8.25	6.29	CNR1	1.57	9.13
S100A8	4.19	2.55	TIMP3	1.56	2.4
IGFBP1	3.96	152	CCL3	1.56	1.68
PROM1	3.70	−1.14	GZMB	1.51	1.81
LGR5	3.64	1.35	SPP1	1.50	1.93
MMP3	3.50	12.62	SERPINA5	−1.59	−1.4
IGF2|INS-IGF2	3.11	2.06	PAPPA|PAPPAS	−1.61	−3.1
TNFRSF11B	3.05	6.04	GPX2	−1.63	−4.49
PRL	2.99	3.38	ICA1	−1.75	−2.41
ITGB3	2.55	4.57	CTNNA2	−1.87	−2.38
MMP1	2.53	52.27	PLCB1	−1.90	−2.37
LEFTY2	2.39	7.33	TLR3	−1.93	−4.61
FMOD	2.26	3.12	FN1	−2.07	−1.19
TREM1	2.25	10.39	TLR4	−2.08	−6.96
RGS2	2.22	2.07	HLA-DOB|TAP2	−2.14	−11.77
FCGR3A	2.13	3.29	RHOU	−2.19	−1.97
C3	2.10	−1.03	GNG11	−2.20	−5.24
WNT5A	1.97	1.82	PLCB4	−2.27	−4.06
FCER1G	1.96	1.52	CYP26A1	−2.88	−4.95
CD163	1.84	2.72	FGF10	−2.91	−11.07
WNT7A	1.84	5.21	OGN	−3.16	−8.03
CCL4	1.82	1.69	TFPI2	−3.59	−2.3
CCL3L1|CCL3L3	1.79	3.42	HGD	−4.87	−9.78
SERPINE1	1.72	3.94	HSD17B2	−5.63	−24.43
PRF1	1.69	1.37	MMP26	−7.29	−12.33

Abbreviations: q-RT-PCR, quantitative reverse transcriptase polymerase chain reaction; DMPA, depot medroxyprogesterone acetate.

^aSpearman’s ρ: 0.82, P < .0001.

^bKendall τ: 0.62, P < .0001.

Table 3.

Correlation of Array and q-RT-PCR: Endometrium LNG-IUS Versus Control.^a,b

Gene	Array	qRT-PCR	Gene	Array	qRT-PCR
IGFBP1	12.33	495	CD97	2.24	2.44
MMP7	11.72	13.5	MMP3	2.22	18.75
PRL	10.22	19.7	CD180	2.19	1.71
S100A8	10.15	20.62	MMP1	2.16	24.48
TREM1	9.57	25.52	CXCR1	2.14	12.4
SST	6.26	57	PLAU	2.11	2.21
SERPINE1	6.23	15.11	CCL7	2.07	123.2
AQP9	6.20	38.6	CCRL2	2.04	4.56
CNR1	5.55	76.7	CCL24	1.88	69.38
CD163	5.15	4.75	PROM1	1.88	1.26
HSD11B1	4.71	132.5	VEGFA	1.79	1.99
FCGR3A	4.32	4.42	MMP2	1.79	2.04
CCL3L1|CCL3L3	4.06	5.43	SERPING1	1.65	1.27
CCL4	3.99	3.42	NFKB1	1.62	1.38
CCL3	3.95	5.42	WNT7A	1.53	7.26
RGS2	3.89	4.64	GPX2	−1.57	−2.64
FCER1G	3.61	3.24	S100A14	−1.62	−1.65
MARCO	3.43	741	RHOU	−1.65	−1.62
CCR1	3.13	3.03	LGR5	−1.90	−2.7
IGF2|INS-IGF2	3.04	2.04	HLA-DOB|TAP2	−2.06	−3.26
FMOD	2.99	3.9	TLR3	−2.15	−2.58
TIMP3	2.95	2	GNG11	−2.21	−2.43
ITGAM	2.95	4.09	ICA1	−2.36	−3.27
CD86	2.90	9.67	OGN	−2.38	−4.83
C3	2.87	1.22	CTNNA2	−2.60	−2.88
GZMB	2.77	2.01	PLCB1	−2.64	−3.83
ITGB3	2.70	2.95	CYP26A1	−2.68	−2.56
PAPPA|PAPPAS	2.64	2.75	PLCB4	−2.72	−4.28
SPP1	2.54	1.86	FGF10	−2.75	−5.66
TNFRSF11B	2.46	10.08	SERPINA5	−3.56	−7.82
WNT5A	2.46	3.08	TFPI2	−3.91	−2.11
CD300A	2.38	3.59	HSD17B2	−4.78	−12.73
PRF1	2.32	3.06	MMP26	−7.07	−17.02
LEFTY2	2.27	4.06	HGD	−9.30	−10.8
MIR223	2.25	20.34

Abbreviations: q-RT-PCR, quantitative reverse transcriptase polymerase chain reaction; LNG-IUS, levonorgestrel intrauterine system.

^aSpearman’s ρ: 0.79, P < .0001.

^bKendall τ: 0.62, P < .0001.

Pathway Analysis: Effects of DMPA

In the endometrium of DMPA users, IPA predicted increased movement of myeloid cells (z = 3.40), adhesion of immune cells (z = 2.35), and inflammatory response (z = 3.68), among other pathways (Table 4). Upstream regulators include TWIST1 (z = 2.00), tumor necrosis factor (TNF; z = 1.62), nuclear factor κ-B (NFκB) complex (z = 1.41), and interleukin 1 receptor antagonist (IL1RN; z = −2.00; Table 5).

Table 4.

Biofunctions: Endometrium DMPA and LNG-IUS Versus Control.

	# Molecules	Z score	Predicted Activity
DMPA versus control
Cell movement
Chemotaxis of leukocytes	12	3.68	Increased
Cell movement of myeloid cells	14	3.40	Increased
Cell movement of phagocytes	14	3.38	Increased
Cell movement of neutrophils	8	2.63	Increased
Cell movement of granulocytes	10	3.21	Increased
Cell movement of mononuclear leukocytes	11	2.00	Increased
Cell-to-cell signaling and interaction
Adhesion of immune cells	11	2.35	Increased
Activation of leukocytes	14	1.19	–
Molecular transport
Quantity of Ca2+	10	1.65	–
Flux of Ca2+	7	1.84	–
Metabolism of reactive oxygen species	8	2.01	Increased
Production of reactive oxygen species	6	1.80	–
Inflammatory response
Inflammatory response	14	3.68	Increased
Tissue development
Vasculogenesis	12	2.01	Increased
Proliferation of endothelial cells	10	1.99	–
Development of epithelial tissue	11	2.08	Increased
LNG-IUS versus control
Cell movement	89	2.15	Increased
Homing of cells	41	3.35	Increased
Cell movement of leukocytes	49	3.06	Increased
Cell movement of phagocytes	32	3.57	Increased
Cell movement of myeloid cells	30	3.56	Increased
Cell movement of granulocytes	19	3.11	Increased
Cell movement of monocytes	17	2.27	Increased
Cell movement of neutrophils	16	3.73	Increased
Cell-to-cell signaling and interaction
Adhesion of immune cells	26	2.97	Increased
Binding of phagocytes	11	2.88	Increased
Binding of T lymphocytes	5	2.24	Increased
Activation of cells	38	1.62	–
Cellular growth and proliferation
Cell death of connective tissue cells	13	−1.85	–
Free radical scavenging
Synthesis of reactive oxygen species	14	2.20	Increased
Metabolism of reactive oxygen species	15	2.56	Increased
Molecular transport
Quantity of Ca2+	17	2.08	Increased
Accumulation of cyclic AMP	7	−2.25	Decreased
Cellular growth and proliferation
Proliferation of endothelial cells	17	2.07	Increased
Cardiovascular system development and function
Development of blood vessel	29	2.02	Increased
Cellular function and maintenance
Engulfment of cells	14	2.02	Increased
Tissue development
Development of epithelial tissue	21	1.67	–

Abbreviations: DMPA, depot medroxyprogesterone acetate; LNG-IUS, levonorgestrel intrauterine system; AMP, adenosine monophosphate.

Table 5.

Upstream Regulators: Endometrium DMPA and LNG-IUS Versus Control.

Upstream Regulator	Molecule Type	# Molecules	Z Score	Predicted Activity
DMPA vs control
TWIST1	Transcription regulator	4	2.00	Activated
TNF	Cytokine	19	1.62	–
NFκB (complex)	Complex	6	1.41	–
IL10	Cytokine	7	1.34	–
IL18	Cytokine	4	1.30	–
LDL	Complex	4	1.29	–
P38 MAPK	Group	6	1.01	–
Immunoglobulin	Complex	4	1.00	–
IL13	Cytokine	9	−1.67	–
IL1RN	Cytokine	4	−2.00	Inhibited
LNG-IUS vs control
TNF	Cytokine	37	3.93	Activated
NFκB (complex)	Complex	16	3.54	Activated
P38 MAPK	Group	14	3.26	Activated
IL1B	Cytokine	19	2.87	Activated
ERK1/2	Group	8	2.80	Activated
IFNA2	Cytokine	7	2.62	Activated
IL2	Cytokine	7	2.60	Activated
TLR7	Transmembrane receptor	7	2.60	Activated
IFNG	Cytokine	21	2.47	Activated
TWIST1	Transcription regulator	6	2.45	Activated
Jnk	Group	6	2.43	Activated
RNASE1*	Enzyme	5	2.22	Activated
IL15	Cytokine	5	2.22	Activated
MIF	Cytokine	5	2.19	Activated
TREM1*	Transmembrane receptor	17	2.18	Activated
RNASE2	Enzyme	8	2.15	Activated
IL1A	Cytokine	4	2.00	Activated
TSLP	Cytokine	4	1.98	–
CSF2	Cytokine	8	1.97	–
CD40LG	Cytokine	4	1.97	–
CTGF	Growth factor	4	1.97	–
EGFR	Kinase	4	1.96	–
TCR	Complex	14	1.95	–
CCL5	Cytokine	7	1.89	–
IL17A	Cytokine	6	1.75	–
TLR2	Transmembrane receptor	6	1.66	–
IL18	Cytokine	8	1.62	–
Cg	Complex	9	1.59	–
STAT3	Transcription regulator	11	1.58	–
IL4	Cytokine	10	1.53	–
Immunoglobulin	Complex	8	1.41	–
IL13	Cytokine	22	1.41	–
LDL	Complex	6	1.38	–
TGFB1	Growth factor	13	1.35	–
IL6	Cytokine	5	1.34	–
IL10	Cytokine	18	1.18	–
IL21	Cytokine	8	1.13	–
IL3	Cytokine	4	1.07	–
Hsp27	Group	4	1.00	–
CD3	Complex	8	−1.07	–
miR-155-5p	Mature microrna	4	−1.98	–
COL18A1	Other	9	−2.36	Inhibited
IL1RN	Cytokine	7	−2.65	Inhibited

Abbreviations: DMPA, depot medroxyprogesterone acetate; LNG-IUS, levonorgestrel intrauterine system.

Pathway Analysis: Effects of LNG-IUS

In the endometrium of LNG-IUS users, pathway analysis suggested an increase in cell movement of myeloid cells (z > 3.00), engulfment of cells (z = 2.02), and binding of phagocytes (z = 2.88) and T lymphocytes (z = 2.24). Pathways related to blood vessel development (z = 2.02) and synthesis and metabolism of reactive oxygen species (z = 2.20, 2.56) were also activated (Table 4). Upstream regulators that were significantly activated included TNF (z = 3.93), NFκB complex (z = 3.54), P38 MAPK (z = 3.26), and IL1B (z = 2.87) among several others (Table 5).

Validation and Correlation Between Microarray and q-RT-PCR

Differentially expressed genes of interest from endometrium (52 DMPA vs control, 69 LNG-IUS vs control) were chosen and Fluidigm DELTAgene assays were performed starting from the same total RNAs used for the microarray. For each comparison, average fold changes were calculated and correlation analyses were performed using 2 nonparametric measures of correlation, Spearman ρ, and Kendall τ, for DMPA versus control (ρ = 0.82, P < .0001; τ = 0.62, P < .0001) and LNG-IUS versus control (ρ = 0.79, P < .0001; τ = 0.62, P < .0001; Tables 2 and 3).

Effects of DMPA and LNG-IUS on Cervical TZ

The cervical transformation is an important site of cell-mediated immunity.²¹ When compared with unexposed controls, there were 31 genes differentially expressed with DMPA and 23 genes in LNG-IUS compared with unexposed controls. Between the 2 treatments, there were 21 genes that shared changes in expression, and 10 and 2 genes whose expression was unique to DMPA and LNG-IUS, respectively (Figure 3). In DMPA users, the GABA_A receptor pi (GABRP) and 15-hydroxyprostaglandin dehydrogenase [NAD+] (HPGD) were the genes that showed the greatest increase in expression, while MMP7 and complement component 3 (C3) were the genes that showed the largest decline in expression. With LNG-IUS, HPGD had the greatest increase in expression, followed by COX7B and GABRP, while C3 and prominin 1 (PROM1) were the most downregulated. Full lists of differentially expressed genes are given in Supplemental Table 2C and D, and those with q-RT-PCR-validated expression levels are given in Table 6.

Table 6.

Correlation of Array and q-RT-PCR: Cervix DMPA and LNG-IUS Versus Control.^a,b

DMPA Versus Control			LNG-IUS Versus Control
Gene	Array	qRT-PCR	Gene	Array	qRT-PCR
GABRP	3.03	4.04	HPGD	2.14	5.27
HPGD	2.44	2.34	COX7B	1.69	1.31
S100A14	1.79	1.13	GABRP	1.69	1.47
CLDN8	1.58	1.94	S100A14	1.66	1.29
SERPING1	−1.55	−1.22	CLDN8	1.54	2.78
TIMP1	−1.61	−1.14	PTGER2	−1.62	−1.02
LOX	−1.63	−2.25	FN1	−1.64	−1.39
CFTR	−1.63	−1.22	SPP1	−1.75	−1.21
HGD	−1.64	−1.34	IL1R1	−1.77	−1.03
FMOD	−1.65	−1.55	MMP2	−1.77	−1.44
IL1R1	−1.70	−1.72	SERPING1	−1.84	−1.15
SPP1	−1.78	−1.53	C3	−1.94	1.20
FN1	−1.78	−1.68	PROM1	−2.09	1.63
MMP2	−1.87	−1.57
C3	−1.93	−1.02
MMP7	−1.93	−1.24
PROM1	−2.42	−1.03

Abbreviations: q-RT-PCR, quantitative reverse transcriptase polymerase chain reaction; LNG-IUS, levonorgestrel intrauterine system; DMPA, depot medroxyprogesterone acetate.

^aSpearman’s ρ: 0.49, P = .48; Spearman’s ρ: 0.46, P = .11.

^bKendall τ: 0.38, P = .039; Kendall τ: 0.33, P = .11.

Pathway Analysis: Effects of DMPA

In the cervical TZ of DMPA users, pathway analysis indicates the occurrence of increased necrosis (z = 2.04) and decreased proliferation of cells (z = −2.30). A trend toward decreased expression of genes involved with adhesion of blood cells (z = −1.20) was also found (Table 7). A trend toward the decreased expression of TNF genes (z = −0.17) was identified but was not statistically significant (Table 8).

Table 7.

Biofunctions: Cervix DMPA and LNG-IUS Versus Control.

	# Molecules	Z Score	Predicted Activity
DMPA vs control
Cell death
Necrosis	5	2.04	Increased
Apoptosis	5	1.31	–
Cell death	6	1.30	–
Cell-to-cell signaling and interaction
Adhesion of blood cells	4	−1.20	–
Cellular growth and proliferation
Proliferation of cells	9	−2.30	Decreased
LNG-IUS vs control
Cellular movement
Migration of cells	6	−1.01	–
Cell-to-cell signaling and interaction
Adhesion of blood cells	4	−1.44	–
Cellular growth and proliferation
Proliferation of cells	7	−2.42	Decreased

Abbreviations: DMPA, depot medroxyprogesterone acetate; LNG-IUS, levonorgestrel intrauterine system.

Table 8.

Upstream Regulators: Cervix DMPA and LNG-IUS Versus Control.

Upstream Regulator	Molecule Type	# Molecules	Z Score	Predicted Activity
DMPA vs control
TNF	cytokine	6	−0.17	–
LNG-IUS vs control
TNF	cytokine	4	−1.15	–

Abbreviations: DMPA, depot medroxyprogesterone acetate; LNG-IUS, levonorgestrel intrauterine system.

Pathway Analysis: Effects of LNG-IUS

In TZ of LNG-IUS users, the expression of genes involved in the proliferation of cells was significantly decreased (z = −2.42), and there was a trend toward decreased expression of genes involved in the adhesion of blood cells (z = −1.44) and migration of cells (z = −1.01; Table 7). As with DMPA, a trend indicating a decrease in genes involved with TNF was identified, but the effect did not reach statistical significance (z = −1.15; Table 8).

Validation and Correlation Between Microarray and q-RT-PCR

Differentially expressed genes of interest from cervix (17 DMPA vs control, 13 LNG-IUS vs control) were chosen and Fluidigm DELTAgene assays were performed starting from the same total RNA extracts used for the microarray. Average fold changes were calculated and correlation analyses were performed using Spearman ρ and Kendall τ tests, for DMPA versus control (ρ = 0.49, P = .048; τ = 0.38, P = .039) and LNG-IUS versus control (ρ = 0.46, P = .11; τ = 0.33, P = .11; Table 6).

Endometrium Versus Cervix: Differential Expression of Genes and Pathway Analyses

The current study also offered the opportunity to study the biology of the TZ that was exposed to different progestins and compare its response to that of the endometrium. Analyses of endometrium versus cervix for DMPA, LNG-IUS, or no treatment were performed. There were 384 differentially expressed genes for no treatment, 515 for DMPA, and 643 for LNG-IUS. In all, 106 genes were common to all 3 groups, while DMPA and LNG-IUS shared 237, DMPA shared 52 with control, and LNG-IUS shared 35 with control (Figure 3). In all analyses, IGFBP1 was upregulated in endometrium. For DMPA and LNG-IUS, PRL, osteopontin (SPP1), fibronectin 1 (FN1), MMP7, and TIMP3 were also among the most upregulated genes in endometrium (Supplemental Table 2E-G). Genes related to cell movement of leukocytes, inflammatory response, and cell death were modified in endometrium of women using DMPA and LNG-IUS but not in the no-contraceptive group. In cervix, compared with endometrium, pathways related to infectious disease including replication of HIV-1 (z = −2.26 for DMPA, −1.64 for LNG-IUS) and infection of T-lymphocytes (z = −2.05, −1.41) were decreased as was cell death of connective tissue cells (z = −1.34, −1.34; Supplemental Table 3). Several upstream regulators were also found to be altered in each treatment (Supplemental Table 4).

Discussion

Progestins Alter Immune Milieu

Our findings herein demonstrate that administration of the progestin-based contraceptives DMPA and LNG-IUS alters the immune milieu of the endometrium, with key pathways governing recruitment of immune cells and immune response being differentially expressed. The pathways we have identified are consistent with studies demonstrating chemokine alterations and increased numbers of leukocytes, including macrophages, in endometrium of subdermal LNG, LNG-IUS, MPA, or norethisterone users.^22,23 The current study comprises the first global gene expression profile of upper FRT tissues in women using progestin-based contraceptives, and the accompanying analysis provides insight into possible mechanisms for the increased rates of HIV infection observed for users of progestin contraceptives in some studies.

Although the cervical TZ is commonly considered to be vulnerable to HIV,^21,24 an appreciable change in immunologically or structurally relevant gene expression in response to progestins was not detected in our work. In contrast, both progestins induced profound changes in endometrial tissue, influencing the expression of cytokines and chemokines, markers of monocyte-derived cells, pattern recognition receptors, and components of the complement system. Together the data indicate that DMPA and LNG-IUS use may increase susceptibility to HIV infection via the recruitment of HIV target cells and a reduced antiviral response by these cells in the endometrium.

Global and Progestational Effects of LNG-IUS and DMPA on Endometrium

In this study, the changes in expression of several genes were consistent with the expected action of both progestins on the endometrium. The high concentration of LNG released directly into the uterine cavity, as well as mechanical effects of the IUS contacting the uterus, may have resulted in the larger number of genes that demonstrated differential expression and the greater extent of the changes in expression found in endometrial tissues compared with tissues taken from users of DMPA, which is systemically and not locally administered. Upregulation of IGFBP1 and PRL, key markers of endometrial stromal decidualization,^25–27 is consistent with their upregulation in stromal fibroblasts exposed to progestins.²⁸ There was also marked downregulation of 17β-hydroxysteroid dehydrogenase type 2 (HSD17B2), which converts estradiol E₂ to the less active estrone E₁, an effect previously shown with progestin use.²⁹

Interestingly, a study published by another group analyzed the effects of an inert intrauterine device on endometrial gene expression and reported a profile very different from our own.³⁰ Although several genes did overlap between the 2 studies, the direction of the change was not necessarily the same. These differences may indicate that the local hormonal effect overrides, or interacts with, the effect of the foreign substance in the uterus.

Depot Medroxyprogesterone Acetate and LNG-IUS Induce Expression of Genes Associated With Leukocyte Targeting to the Endometrium

Although CD4 T cells are a primary target in productive HIV infection, antigen presenting cells (APCs) that are concentrated at vulnerable mucosal sites are believed to be initial sites of viral replication.^31,32 Dendritic cells (DCs) populate the cervicovaginal squamous epithelium as well as intraepithelial and submucosal sites of the endocervix and endometrium.^33,34 In many cases, these cells extend processes through columnar epithelium to probe for foreign antigen.³⁵ Thus, they may be the first point of contact for HIV and are central to viral dissemination to the underlying tissue. Both macrophages and DC may be infected by HIV at low frequencies and even in the absence of infection can transfer virus to T cells through direct contact,^36,37 a process aided by expression of CD4, CCR5, and other HIV-binding receptors on both cell types.^33,38 In the current study, progestin users demonstrated pathways related to increased myeloid cell migration and binding in the endometrium, as well as increased expression of myeloid cellular markers, lending support to the idea that these cells may contribute to an increased risk of HIV acquisition in progestin users.

During the secretory phase of normally cycling endometrium, an increase in uterine natural killer cells, macrophages, mature DC, and neutrophils, but not B or T lymphocytes, has been observed.³⁹ In the current study, for both DMPA and LNG-IUS compared with mid-secretory phase unexposed controls, changes in expression of genes involved in lymphocyte chemotactic pathways were not found, with differentially expressed genes predominantly predicted to increase myeloid cell chemotaxis and adhesion (Table 4). In particular, there was an increase in the expression of macrophage and DC markers, including CD14 and CD163, suggesting increased numbers of these cells in the endometrium of progestin users.

Interestingly, in the DMPA versus control analysis, inflammatory response was one of the upregulated pathways. Although the effects of progestins are widely thought to be anti-inflammatory, it is important to note the context in which these responses occur. During the progesterone-mediated secretory phase, a variety of leukocytes are recruited to the endometrium during the window of implantation by endometrial epithelial, stromal, and resident leukocyte-secreted chemokines and cytokines, such as interleukin (IL) 6, IL-8, and tumor necrosis factor alpha (TNFA).⁴⁰ These cells participate in a variety of functions: Decidual NK cells secrete angiogenic factors necessary for a vascularized decidua.⁴¹ Studies in mice have shown that endometrial DCs promote a cytokine profile in endometrium that is necessary for successful decidualization.⁴² Both cells are major facilitators of the innate immune response, and studies (including our results) indicating that progestins promote an inflammatory, chemotactic effect is consistent with events occurring in vivo. Th2 T regulatory cells, which suppress the immune system to promote immunotolerance of the implanting embryo, are also recruited to endometrium, and hence the “anti-inflammatory” response that has been classically proposed as the effect of progesterone/progestins on the endometrium.⁴³ Dendritic cells, in addition to decidualization, have also been shown to play a part in this type of immunosuppression.⁴⁴ That immune cells are recruited to the endometrium during the secretory phase, mimicking an inflammatory response to infection and that the milieu that the recruited cells generate is anti-inflammatory (because those cells suppress the immune system to inhibit allogenic responses to the implanting embryo) underscore interpretation of the type of inflammatory response elicited by progesterone and synthetic progestins. Furthermore, progestins may elicit cellular responses that differ from progesterone per se and contribute to the results observed herein.

Genes Involved in Microbial Recognition and Antiviral Response are Altered in Endometrium of DMPA and LNG-IUS Users

Pattern recognition receptors such as C-type lectins, scavenger receptors, CD14, and toll-like receptors (TLRs) are crucial mediators of antimicrobial responses in APCs. Strikingly, genes for many of these receptors were differentially expressed in response to progestins (Table 9), including decreased TLR3 with both progestins, downregulation of TLR4 with DMPA, and upregulation of TLR8 with LNG-IUS. Expression of the HIV-enhancing DC-SIGN (CD209) and mannose receptor (MRC1; CD206), expressed on dendritic cells or macrophages residing within the uterine luminal epithelium,³⁴ was also increased with LNG-IUS.

Table 9.

Select Differentially Expressed Genes Listed by Category.

Gene	Regulation	Fold change (array)
		DMPA	LNG-IUS
Pattern recognition receptors
CD14	Up	1.86	3.81
CLEC12A	Up		1.54
CLEC2D	Up	2.06	2.57
DC-SIGN	Up		1.56
MARCO	Up		3.43
MRC1|MRC1L1	Up		3.83
MSR1	Up	2.19	3.15
SCARA3	Up	1.70
TLR3	Down	−1.93	−2.15
TLR4	Down	−2.08
TLR8	Up		1.76
Complement components
C2 | CFB	Up		1.53
C3	Up	2.10	2.87
CR1	Up		2.23
ITGAM (CR3)	Up		2.95
ITGAX (CR4)	Up	1.78	3.25
SERPING1	Up		1.65
Metallopeptidases and inhibitors
ADAM19	Up		1.68
ADAMTS2	Up		1.76
ADAMTS5	Up		1.74
ADAMTS8	Down	−1.85	−1.62
MMP1	Up	2.53	2.16
MMP16	Down		−1.57
MMP19	Up		2.26
MMP2	Up		1.79
MMP26	Down	−7.29	−7.07
MMP3	Up	3.50	2.22
MMP7	Up	8.25	11.72
TIMP1	Up		2.69
TIMP3	Up	1.56	2.95
TIMP4	Up		1.56
Cytokines and receptors
CCL24	Up		1.88
CCL3	Up	1.56	3.95
CCL3L1 | CCL3L3	Up	1.79	4.06
CCL4	Up	1.82	3.99
CCL7	Up		2.07
CCL8	Up		2.47
CCR1	Up		3.13
CCRL2	Up		2.04
CMKLR1	Up		1.93
CSF1	Up		1.71
CSF2RA	Up		1.67
CXCL3	Up		1.52
CXCR1	Up		2.14
IL2RA	Up		1.58
TNFRSF11B	Up	3.05	2.46

Abbreviations: DMPA, depot medroxyprogesterone acetate; LNG-IUS, levonorgestrel intrauterine system.

The complement system is an important mediator of innate and adaptive immune responses, activation of which leads to attraction of macrophages and neutrophils and ultimately lysis or phagocytosis of pathogens. Work by our group and others has shown that expression of complement components in the endometrium is regulated by steroid hormones.^9,45 Interestingly, C3 and the α integrin component of CR4 (ITGAX; CD11c) were upregulated with DMPA and LNG-IUS, herein, and CR1, CR2, C1 inhibitor (SERPING1), and the α integrin of CR3 (ITGAM; CD11b) were upregulated with LNG-IUS. Although the biological implications of complement component expression in women using progestin contraceptives are not yet fully understood, these data raise important issues about HIV uptake and systemic spread via lymphoid tissue.

Genes Regulating Vascular Development are Differentially Expressed in Endometrium of DMPA and LNG-IUS Users

With both progestins, there were changes in multiple genes predicted by IPA to increase blood vessel development and endothelial cell proliferation (Table 4), including several MMPs, members of the ADAMTS family, tissue inhibitors of metalloproteinases (TIMPs), and SERPINE1 (endothelial plasminogen activator inhibitor, PAI-1). With LNG-IUS, levels of vascular endothelial growth factor A (VEGFA) also increased. Some of these changes have been previously described and are associated with abnormal bleeding in progestin users.^46–48 In combination with other pathways described herein, increased vascularization in the endometrium is a potential route of virus dissemination to the general circulation.

Global and Progestogenic Effects of LNG-IUS and DMPA on Cervix

In cervical TZ, more genes and higher fold changes were observed with DMPA than LNG-IUS. Local release of LNG in the endometrium may explain the lesser effect in the cervix, whereas systemic DMPA could access this tissue more readily. Nonetheless, gene expression changes overlapped considerably between DMPA and LNG-IUS, attesting to their similar mechanisms of action. Interestingly, GABA_A receptor pi (GABRP) was among the most highly upregulated genes after exposure to either progestin. Differential expression of genes related to tissue structure is reminiscent of changes occurring during pregnancy with progesterone treatment,⁴⁹ cervical softening and ripening.^50,51 In DMPA users, genes for lysyl oxidase (LOX), fibromodulin (FMOD), FN, SPP1, and MMP2 and MMP-7 were downregulated, suggesting coordinated tissue reorganization. Some of these same genes were downregulated with LNG-IUS as well. Indeed, progesterone and progestins have been used to effectively prevent preterm labor in some populations,^52–54 possibly by inhibiting cervical ripening,⁵⁵ a process characterized by extracellular matrix reorganization. Few human studies have looked at the role of steroid hormones in cervical ripening biochemically; however, expression of estrogen receptor and progesterone receptor correlate with collagen and proteoglycan composition between nonpregnant, term and postpartum cervix.⁵⁶ Collagen synthesis by cultured human cervical fibroblasts is decreased in response to progesterone, an effect partially abrogated by mifepristone.⁵⁷ Overall, these data support a transcriptional response in the cervix of DMPA and LNG-IUS users, and future studies should aim to better define these changes in the context of HIV.

Gene Expression in Response to DMPA and LNG-IUS Suggests Reduced Proliferation and Viability of Cells in Cervical TZ

In contrast to effects observed in endometrium, exposure to DMPA or LNG-IUS had little effect on the transcription of immune-related genes in the cervix, although reduced expression of genes implicated in the adhesion of blood cells was suggested for both. Pathways related to cellular proliferation were decreased and in DMPA pathways involved in tissue necrosis were increased; however, these predictions were made based on fairly small changes in the expression of a small number of genes. In addition, expression of HPGD was upregulated in response to both DMPA and LNG-IUS. Under hormonal control,^58,59 HPGD degrades prostaglandin E2, a vasoactive compound that attracts immune cells to the cervix during ripening.^60,61 Thus, HPGD upregulation is consistent with leukocyte efflux from the cervix. A recent study found that, in contrast to the leukocyte influx that occurs in labor,^62,63 MPA reduces macrophage and neutrophil populations in the pregnant mouse cervix.⁵⁵ Based on these data, it is unlikely that an influx of HIV target cells to the cervix can explain increased susceptibility to HIV following progestin use. However, this does not preclude changes in localization, activation, or viability of cells within the cervix. Given the regulation of genes related to tissue remodeling of the cervical TZ, these other mechanisms are worthy of further investigation.

Progestins Induce Pathways of Tissue Remodeling and Immune Activation in Endometrium Compared With Cervix

In DMPA or LNG-IUS users, pathways regulating cell death, proliferation, and motility and adhesion of leukocytes were found to be increased in endometrium compared with cervix. For women not using progestins, fewer differences were found between the 2 tissues. This can be seen in the PCA, as control samples from each tissue cluster more closely to each other. These data suggest that endometrium but not cervix responds profoundly to both progestins. Interestingly, expression of genes in pathways related to viral infection including replication of HIV-1 was decreased in the endometrium of controls, based on expression of cytokines such as CCL3 and CCL4 in both progestins and TLR-4 and -8 with DMPA. These findings indicate that cervical TZ has a limited transcriptional response to progestins, and in contrast with endometrium, cervical tissue may be more susceptible to HIV in the absence of exogenous progestins.

Limitations to Interpretation of the Study

This study provides a global transcriptomic analysis of endometrium and cervix in vivo in women using progestin contraceptives. Although a minimum of 6 months progestin use was required for participation in the study, due to a small sample size, we did not analyze the effects of exposure time on gene expression; however, we think it is important to investigate this variable in the future.

Given that this was a whole tissue study, it is difficult to draw conclusions about cell-specific responses in these tissues. To address this, we have begun in vitro studies whereby isolated cells are treated with biologically relevant levels of both MPA and LNG to elucidate cell-specific responses relevant to HIV infection. In addition, we are in the process of validating immunological changes to upper FRT tissues, including populations of HIV-target cells, using flow cytometry, immunohistochemistry, and protein quantification.

Summary

By performing microarray analysis on cervical and endometrial tissues of women using the progestin-based contraceptives DMPA and LNG-IUS, we have provided a global assessment of progestin-induced and repressed gene expression changes in the upper FRT. In addition, we have identified potential immunological alterations in the endometrium that are pertinent to known modes of mucosal HIV infection, including recruitment of HIV-susceptible cells, dampening of innate antiviral mechanisms, and increased vascularization of the tissue (Figure 4). In the cervical TZ, genes related to tissue reorganization were modified in a stereotyped manner, although few changes in the immune milieu were observed. Together, these findings lead us to emphasize the need for further examination of progestin-induced responses in upper FRT tissues in the context of HIV infection.

Figure 4.

Summary and implications of pathways regulated by progestins in endometrium and transformation zone (TZ).