Abstract

   Preterm delivery is both a traumatizing experience for the patient and
   a burden on the healthcare system. A condition distinguishable by its
   phenotype in prematurity is cervical insufficiency, where certain cases
   exhibit a strong genetic component. Despite genomic advancements,
   little is known about the genetics of human cervix remodeling during
   pregnancy. Using selected gene approaches, a few studies have
   demonstrated an association of common gene variants with cervical
   insufficiency. However, until now, no study has employed comprehensive
   methods to investigate this important subject matter. In this study, we
   asked: i) are there genes reliably linked to cervical insufficiency
   and, if so, what are their roles? and ii) what is the proportion of
   cases of non-syndromic cervical insufficiency attributable to these
   genetic variations? We performed next-generation sequencing on 21
   patients with a clinical presentation of cervical insufficiency. To
   assist the sequencing data interpretation, we retrieved all known genes
   implicated in cervical functioning through a systematic literature
   analysis and additional gene searches. These genes were then classified
   according to their relation to the questions being posed by the study.
   Patients’ sequence variants were filtered for pathogenicity and
   assigned a likelihood of being contributive to phenotype development.
   Gene extraction and analysis revealed 12 genes primarily linked to
   cervical insufficiency, the majority of which are known to cause
   collagenopathies. Ten patients carried disruptive variants potentially
   contributive to the development of non-syndromic cervical
   insufficiency. Pathway enrichment analysis of variant genes from our
   cohort revealed an increased variation burden in genes playing roles in
   tissue mechanical and biomechanical properties, i.e. collagen
   biosynthesis and cell-extracellular matrix communications.
   Consequently, the proposed idea of cervical insufficiency being a
   subtle form of collagenopathy, now strengthened by our genetic
   findings, might open up new opportunities for improved patient
   evaluation and management.

Introduction

   In order to carry a successful term pregnancy, different organs such as
   the uterus, cervix, placenta, and amniotic membranes as well as the
   fetus itself must cohesively interact and create a healthy symbiotic
   relationship with each other and the rest of the female body [[37]1].
   However, preterm birth (PTB) remains the leading cause of perinatal
   morbidity, mortality, and hospitalization in the first year of life in
   the developed world. Approximately 5–12% of newborns worldwide are born
   preterm (<37 weeks of gestation) [[38]2]. Prematurity is a tremendous
   burden on the healthcare system as outcomes are associated with
   disability-specific lifetime medical, special education, and lost
   productivity costs [[39]3].

   A common phenotype of spontaneous PTB is primarily characterized by
   progressive cervical effacement, after which preterm premature rupture
   of membranes (PPROM), persistent uterine contractions, prolapsed fetal
   membranes, or uterine bleeding may be the reason for acute care
   seeking.

Isolated cervical insufficiency

   A distinguishable medical condition in obstetrics in which the cervix
   spontaneously starts to dilate (open) and efface (become thinner) in
   the absence of the signs and symptoms of labor is cervical
   insufficiency. The cervix, a collagen-rich organ, must remain closed
   during pregnancy yet simultaneously undergo a progressive physiological
   remodeling to prepare for the birth. Physiological cervical remodeling
   along with uterine contractile activation are the two key events
   facilitating the birth of a child [[40]4]. This remodeling can be
   loosely divided into four overlapping phases: 1) softening beginning in
   early pregnancy, 2) ripening shortly before the birth, 3) dilation
   starting with the onset of regular uterine contractions and resulting
   in cervical opening to allow passage of the term fetus, and 4)
   postpartum repair [[41]4–[42]6]. In cases of cervical insufficiency,
   dilation of the cervix occurs without painful uterine contractions,
   leading to inability of the cervix to retain a term pregnancy.
   Repeating in consecutive pregnancies, cervical insufficiency is one of
   the causes of recurrent pregnancy loss [[43]7] and can be a serious
   obstacle to the birth of a healthy child and complication-free
   postpartum period for the mother and newborn. In contrast, failure of
   the cervix to dilate would result in unsuccessful parturition [[44]8].

   Clinically relevant isolated cervical insufficiency occurs in about
   1–2% of all pregnancies, but is associated with as much as 5–15% of
   pregnancy losses in the second trimester [[45]9,[46]10]. However, as
   one of the factors in a complex PTB context, the condition is found
   much more frequently. In 2011, routine recording of cervical ripening
   was recommended by the Global Alliance to Prevent Prematurity and
   Stillbirth [[47]11], since a short cervix (defined as a transvaginal
   sonographic cervical length ≤25 mm in the mid-trimester of pregnancy)
   is the best predictive factor for spontaneous PTB <34 weeks of
   gestation in both singletons and twins [[48]12]. The shorter the
   cervix, the higher the risk; cervical insufficiency is likely at the
   extreme of this continuum [[49]13].

   Multiple factors such as age, inflammation, stress, nutrition, physical
   activity, socio-economic status, vaginal microbiome, and uterine
   anomalies affect PTB [[50]14–[51]16]. Mid-trimester cervical weakness
   may be associated with a variety of events, e.g. cervical ablation
   (cryo, laser, or electro) or excision (knife, laser, or
   loop-electrosurgical), cervical intraepithelial neoplasia per se,
   cervical hypoplasia after diethylstilbestrol, or intrauterine
   infections [[52]17].

   PTB is currently perceived as a frequent complex medical condition
   which corresponds to the concept of a multifactorial disorder
   [[53]18]–analogous to, for example, cardiovascular disease–the
   development of which depends on a number of interacting factors
   including environmental and genetic. Familial aggregation is evident in
   prematurity [[54]19,[55]20], including cases of cervical insufficiency,
   with up to 27% of patients having a first-degree relative with the same
   diagnosis on the mother’s side [[56]21,[57]22]. By contrast, the risk
   appears to be unaffected by a history of prematurity in the partner’s
   family [[58]23]. Epidemiological data show that fetuses/neonates with
   Ehlers-Danlos syndrome (EDS), osteogenesis imperfecta, and restrictive
   dermopathy are at an increased risk of adverse pregnancy outcomes
   including PTB, PPROM, and cervical insufficiency [[59]24,[60]25]. A few
   studies have demonstrated a positive association of common gene
   variants in the mother’s genome with cervical insufficiency
   [[61]22,[62]26,[63]27].

Current understanding of genetics of cervical remodeling during pregnancy is
limited

   Prior to the era of ‘-omics’, the majority of studies investigating the
   role of genetics in prematurity targeted candidate genes with known
   biological roles potentially related to processes occurring during
   pregnancy [[64]18,[65]28]. For example, common allelic
   variants/polymorphisms in TNF, IL1B and IL6 genes have most
   consistently been associated with PTB [[66]29], underlining the role of
   inflammation in the pathogenesis of prematurity.

   More recent studies on the genetics of PTB in humans can be roughly
   divided into two major categories. The first group is comprised of a
   small number of large-scale genomic studies investigating possible
   genetic risk factors for preterm delivery [[67]30–[68]33].
   Unfortunately, none of these studies has addressed PTB as a result of
   cervical insufficiency. The second group consists of transcriptomic
   studies evaluating differential gene expression during different stages
   of gestation/parturition in eventless gestations [[69]34–[70]36] and
   ones of particular phenotypes, e.g. cases of PTB (HP:0001622) or PPROM
   (HP:0001788; OMIM:610504) [[71]37,[72]38]. For a comprehensive
   evaluation, please refer to the excellent systematic review and
   meta-analysis of [[73]39].

   Although the largest number of studies has focused on idiopathic PTB,
   this phenotype should be considered with caution since preterm delivery
   often encompasses cervical insufficiency, PPROM, placental abruption,
   uterine overdistension, or a combination of these complications
   [[74]13]. This idea is further supported by a meta-analysis of gene
   expression studies across distinct gestational tissues and clinical
   phenotypes which demonstrated a limited overlap of genes identified as
   differentially expressed across the studies [[75]39]. This suggests
   possible different physiological mechanisms underlying each phenotype
   and also indicates that large gaps still exist in the design of
   transcriptomic studies in prematurity.

   Furthermore, in order to attain the true transcriptomic signature of a
   certain phenotype, the tissue of study should be chosen wisely.
   Considering the known heterogeneity of certain tissues, even the biopsy
   site may have an impact on the results. For example, a recent study has
   highlighted that it remains to be resolved whether a PPROM signature
   can be determined in the cervix as the gene expression patterns in
   cervical biopsies of PPROM in comparison to preterm labor samples did
   not share cluster membership, suggesting a distinct genetic signature
   specific to PPROM pathology [[76]37]. Nonetheless, the authors proposed
   the notion that the rupture of membranes might be accelerated through
   PPROM-specific remodeling events within the cervix [[77]37]. Of note,
   cervical insufficiency often results in unscheduled PPROM as well.
   Moreover, similar to the cervix, the mechanical strength of the fetal
   membranes is mainly ensured by the collagen network [[78]40,[79]41].
   Indeed, collagen types I, III, IV, V, and VI, to name but a few, have
   been localized in both cervical (derived from the two paramesonephric
   ducts during embryogenesis) and fetal membrane (derived from the outer
   trophoblast layer of the implanting blastocyst) tissue [[80]42,[81]43].

   Without doubt, our current understanding of human cervix remodeling in
   pregnancy is limited [[82]1]. This may be the reason for the bias of
   studied genes in relation to cervical insufficiency and the
   surprisingly little information that presently exists on the genetics
   of pathological cervical remodeling during pregnancy.

   Since common variants detectable by genome-wide association studies
   (GWAS) typically explain only a minor proportion of the heritability of
   complex diseases [[83]44], there is a hypothesis that the rare variants
   in multiple genes implicated in PTB may cumulatively contribute to the
   predisposition of delivering preterm [[84]15,[85]45]. We decided to
   test this hypothesis by performing next-generation sequencing (NGS) of
   the DNA of females with a positive anamnesis of isolated non-syndromic
   cervical insufficiency.

   Due to the lack of knowledge of genes implicated in cervix functioning,
   we also conducted a systematic literature analysis to derive all
   possible studies on the genetics of the cervix. We subsequently
   composed a list of genes that play a role in the normal and
   pathological biology of the cervix in relation to pregnancy and
   prematurity and used this obtained knowledge to assist our NGS data
   interpretation. Given the described heritability of cervical
   insufficiency, the main questions we addressed in this study were: i)
   are there genes reliably linked to cervical insufficiency and, if so,
   what are their roles? and ii) how many cases of isolated non-syndromic
   cervical insufficiency are attributable to these genetic variations?

Materials and methods

Identification of genes playing a role in the biology of the cervix

Systematic literature analysis: Search strategy and study selection

   We conducted a literature search according to the PRISMA guidelines
   [[86]46] ([87]Fig 1). The screening strategy aimed to retrieve studies
   focusing on genetic research of defective uterine cervix functioning
   leading to cervical insufficiency, preterm delivery, or pregnancy loss,
   as well as records on functional studies addressing the differential
   expression of genes within the cervix during different stages of
   normal/compromised pregnancy/parturition.

Fig 1. PRISMA flowchart depicting the literature search and gene extraction.

   [88]Fig 1
   [89]Open in a new tab

   The search was performed using the MeSH term “Uterine Cervical
   Incompetence” (D002581, which is indexed under two higher categories in
   the MeSH hierarchy: “Pregnancy Complications” and “Female Urogenital
   Diseases”) OR the following keywords: “precocious cervical ripening”,
   “cervical weakness”, “cervical insufficiency”, “istmocervical
   insufficiency”, “cervical incompetence”, “uterine cervix” AND “gene”,
   “genetics”, “gene expression”, “gene transcription”, “transcriptome”
   AND Humans [Mesh] NOT “cancer”, “microbiome”, “papilloma”. The search
   was performed in PubMed, EBSCO Host database, Web of Science, ProQuest
   database, and Primo database. The search was performed in parallel by
   two reviewers on December 16, 2019, without further restrictions on the
   publication date.

   Inclusion criteria: Study published in a peer-reviewed journal; Study
   presents original data; Study concentrates on finding a genetic cause
   of cervical insufficiency and/or preterm delivery; Study concentrates
   on functional gene analysis of physiological cervical ripening,
   cervical insufficiency, and/or preterm delivery as a source using
   cervical tissues. Only human studies were included.

   Exclusion criteria: Study concentrates on miscarriage and/or the first
   trimester of pregnancy; Study concentrates on microRNA, lncRNA,
   cell-free DNA, ribosomal DNA, cervico-vaginal microbiome, cancer
   analysis; Study is not in humans; Study is not available in English
   ([90]S1 Table).

   From the eligible papers presenting original data ([91]S2 Table), we
   extracted the following: gene symbols in a HUGO Gene Nomenclature
   Committee-approved manner; recorded patient phenotypes, i.e. condition,
   or relevancy to our study’s questions if healthy individuals were
   analyzed; biological material used for the analysis (DNA, cervical
   biopsy, other); type of study, e.g. functional, association, or other;
   information on the gene selection approach (unbiased genome-wide or
   selected gene approach). Screening of all the reviews (n = 18) for
   original references did not yield any additional articles to those in