Distinctive structure, composition and biomechanics of collagen fibrils in vaginal wall connective tissues associated with pelvic organ prolapse

Abstract

Collagen is the predominant structural protein within connective tissues. Pelvic organ prolapse (POP) is characterized by weakening of the pelvic floor connective tissues and loss of support for pelvic organs. In this study, we examined the multiscale structure, molecular composition and biomechanics of native collagen fibrils in connective tissues of the posterior vaginal fornix collected from healthy women and POP patients, and established the correlation of these properties with clinical POP quantification (POP-Q) scores. The collagen characteristics, including collagen amount, ratio of Collagen I and Collagen III, collagen fibril D-period, alignment and stiffness, were found to change progressively with the increase of the clinical measurement of Point C, a measure of uterine descent and apical prolapse. The results imply that a severe prolapse is associated with stiffer collagen fibrils, reduced collagen D-period, increased fibril alignment and imbalanced collagen synthesis, degradation and deposition. Additionally, prolapse progression appears to be synchronized with deterioration of the collagen matrix, suggesting that a POP-Q score obtained via a non-invasive clinical test can be potentially used to quantitatively assess collagen abnormality of a patient’s local tissue.

1. Introduction

Pelvic organ prolapse (POP) is a common condition that affects adult women of all ages and compromises their quality of life [1]. In this disorder, pelvic organs protrude through the vaginal inlet due to acquired or inherent abnormalities of the supportive structure of pelvic floor [2]. While pelvic floor muscle plays an important role in visceral support, increasing evidence suggests that abnormalities of pelvic floor connective tissue predispose many women to POP [3,4] and a link between connective tissue disorders and POP has been demonstrated [5].

Pelvic floor connective tissue is composed of collagen, elastin and smooth muscle cells (SMCs) and provides a suspensory system to the pelvic organs [6]. The tissue of POP patients is characterized by weak mechanical strength and compromised biochemical properties, such as decreased amounts of elastin and collagen, altered collagen subtype ratios, increased MMP expression, decreased lisyl oxidase (LOX) expression and reduced number of SMCs [7–10]. Collagen is the most abundant protein in connective tissues [11] and collagen I (COLI) and III (COLIII) are the primary collagen subtypes. While fibrillar COLI offers high tensile strength to tissues, COLIII is commonly found alongside COLI in soft tissues that require increased flexibility and distension [12–14]. Tropocollagen molecules hierarchically assemble into a fibrous matrix that supports and integrates with elastin, SMCs, and other tissue constituents contributing to the mechanical strength of the connective tissues [15,16]. Defects in collagen metabolism and assembly can induce failure of the supportive structure, hence, the loss of tissue strength [17]. Women with collagen-associated disorders tend to have more severe prolapse symptoms and are more likely to have recurrence after surgical repair [18]. Consistent with reports in the literature [7,19,20], reduced collagen content, increased COLI to COLIII ratio and abnormal nanoscopic to microscopic structures of fibrillar collagen in vaginal wall connective tissues of POP patients were observed in our previous studies [21,22]. It remains unclear how these variations are relevant to the severity of prolapse. In this work, we explored the patient-specific, multi-scale studies of collagen in tissues of non-POP and POP individuals to identify features that changed progressively with prolapse severity.

POP severity is conventionally classified into Stages 0-IV by the degree of anatomical deformity on the basis of the most distal portion of prolapse relevant to the level of the hymen. A universal and more reliable POP quantification (POP-Q) system has been established. It employs 6 defined POP-Q points and 3 landmarks (see Figure S1) for objective, site-specific, quantitative measurements [10,23,24]. In this work, we aimed to examine the correlations between individual patient’s clinical scores and the biochemical, biophysical and biomechanical features of collagen in the patient’s tissue to identify a set of strongly correlated parameters. We anticipate that POP-Q scores obtained via a non-invasive clinical test can potentially be used to quantitatively assess collagen abnormality of a patient’s local tissue, allowing clinicians to employ a personalized treatment that can best manage the patient’s condition and to identify novel targets for development of new strategies.

2. Materials and methods

2.1. Clinical tissue biopsy collection and sectioning

With informed consent following the approved IRB protocol at Northshore Health System (IRB protocol #EH15–340), full thickness biopsies, about 2 cm (L) × 0.3 cm (H) × 0.5 cm (W), of posterior vaginal fornix (see Figure S1) were obtained at the time of surgery from 3 non-prolapsed individuals, who had total hysterectomies for reasons other than prolapse, and 14 POP patients who had documented POP-Q scores and no hysterectomy. Tissues harvested at the posterior vaginal fornix were used in this study, as our initial effort focused on studying apical prolapse and the posterior vaginal fornix offers the best access of vaginal apical tissues. Patients with diabetes, gynecologic cancers, and collagen vascular disease were excluded in this study. Subjects were not excluded or categorized based on hormone replacement therapy, history of smoking, or BMI.

All vaginal wall tissue specimen were examined histologically. After blood removal, biopsy samples were cooled to −20 °C for sequential thin sectioning at a thickness of 10 μm or 20 μm using a Shandon Scientific Cryotome (Thermofisher). The sequential thin sectioning of the tissues allowed for adjacent slices to be nearly identical in all aspects. The first and third slices of the tissue were stained with Gomori trichrome to identify collagen from other constituents in the tissue, whereas the slice in between was used in atomic force microscope (AFM) studies at the identified collagen-rich region [21,22]. Additional thin sections were used for immunofluorescent imaging.

2.2. POP-Q measurements

Clinical assessment of POP was based on the established POP-Q system (see Figure S1). Among the POP-Q measurements, Points Aa and Ap are 3 cm proximal to or above the hymenal ring anteriorly and posteriorly, respectively; Points Ba and Bp are defined as the lowest points of the prolapse between Aa anteriorly or Ap posteriorly and the vaginal apex. Anteriorly, the apex is point C (cervix); posteriorly, the apex is point D (pouch of Douglas). Three other measurements are taken: the total vaginal length at rest (TVL), the genital hiatus length (GH) from the middle of the urethral meatus to the posterior hymenal ring, and the perineal body length (PB) from the posterior aspect of the genital hiatus to the mid-anal opening. Each value is in centimeters with the plane of the hymen being defined as zero.

2.3. Gomori trichrome staining

Staining was carried out according to the manufacturer’s protocol. In brief, tissue samples were fixed for 15 mins with 4% paraformaldehyde in phosphate buffered saline (PBS) at pH 7.4 at room temperature, followed by immersion in Bouin’s solution at 56 °C for 1 h. The slides were thoroughly rinsed with DI water prior to applying trichrome stain for 10 mins, then rinsed in acid-alcohol solution for 5 mins and air-dried. The slides were then preserved by progressive rinsing with 95% alcohol (5 mins, twice) and xylene (10 mins) and mounted on a coverslip using Permount mounting medium. Optical images were collected by a Nikon TE-U 2000 microscope (Nikon Instruments Inc., Melville, NY).

2.4. Immunofluorescent imaging

Immunofluorescent imaging of a tissue sample from each patient was carried out following the previously established protocols using the same set of antibodies [22]. Tissue samples were fixed with pure methanol for 5 mins at −10 °C, followed by air-dry and blocking with freshly prepared 10% BSA-PBS for 20 mins. Primary antibodies against COLI (#PA5–95,137) and COLIII (#MA1–22,147), purchased from Thermofisher Scientific (Waltham, MA), were prepared in 1:200 dilutions as recommended by the manufacturer, and were incubated with the tissue on a glass slide for 45 mins in a humidified chamber. Secondary antibodies (#A21206, #A21203), purchased from Invitrogen (Carlsbad, CA), were used at a dilution of 1:200 and incubated with the samples for 1 h in a dark, humidified chamber.

To rule out the influence of dye color in the fluorescence intensity measurement, pure secondary antibodies labeled with Invitrogen^™ Alexa Fluor^™ 488 (#A21206) and Invitrogen^™ Alexa Fluor^™ 594 (#A21203), respectively, were used for dye intensity calibration. Specifically, a drop of each solution, at 1:1000 dilution, was placed on a cover slip to measure intensity by fluorescence imaging under the same imaging conditions as used for the experimental specimen.

Fluorescent images were collected using a Nikon TE-U 2000 microscope. For each experiment on a particular marker, imaging parameters, such as the exposure time, gain value, image size, and magnification, were kept constant through the entire study for all samples. The relative amounts of COLI and COLIII in the connective tissues were analyzed using ImageJ software (National Institute of Health, Bethesda, Marylan) [25,26] based on 15–20 images (20 ×) per patient specimen collected at the center of the specimen’s mucosa region. Background signals were subtracted for fluorescence intensity analysis, and the collagen content was evaluated by fluorescence intensity per unit area of the stained regions.

2.5. AFM imaging and fibril stiffness measurements

Collagen structure and collagen fibril stiffness in a tissue sample were examined using an Agilent Pico-Plus AFM in fluid contact mode. Tissue specimen were pre-treated with ice cold pure methanol for 5 mins followed by air-drying. Measurements were carried out in PBS (pH 7.4) using a closed-loop scanner. Si₃N₄ probes (#SINI-100, Ted Pella, Redding, CA) with a tip radius of less than 15 nm were used and the spring constant of 0.06 ± 0.01 N/m was calibrated using reference cantilevers with known spring constants [21,22,27,28]. Guided by Gomori staining on the adjacent tissue slices, an AFM tip was typically positioned at the center of a collagen-rich mucosa region for high-resolution imaging and force measurement. Fibril width and D-period were measured manually from 7 × 7 μm² images, and the mean values were calculated based on >200 measurements on at least 50 fibrils at 3–5 randomly selected locations for each patient’s specimen.

Young’s modulus, or E-value, of collagen fibrils was derived from the force-distance curves using the Hertzian model with the AFM tip modeled as a nano-indenter to probe the fibrous matrices [29–31]. The mean Young’s modulus of each subject’s tissue was calculated based on a minimum of 1000 force curves collected at 2–4 randomly selected locations at the mucosa region, where densely distributed collagen fibrils were observed (see Fig. 1 and Fig. 3). To make reasonable comparisons between samples from different patients, the indentation depth of all measurements was controlled at 85 ± 10 nm.

Fig. 1.

Multi-scale characterization of vaginal wall tissues. a) Gömöri staining of a 10-μm thick tissue sample illustrating the epithelium, mucosa and muscularis layers. Collagen is stained blue-green. Cells, including smooth muscle cells, are red-brown. Scale bar: 600 μm. b) Immunofluorescence images of COLI (green) and COLIII (red) in the mucosa region of (a). Scale bar: 160 μm. c) AFM image of collagen fibrils at the mucosal region of a 20-μm thick vaginal wall tissue sample. Scale bar: 5 μm. d) High-resolution images of (c) illustrating the fibril width and the characteristic collagen D-period (red bar). Scale bar: 1 μm in (d); 200 nm in inset. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Fig. 3.

AFM images of collagen fibrils in the mucosal region of vaginal wall tissues of individuals with increasing Young’s modulus and patient’s Point C value. (a,d) E = 236 kPa, Point C = −8 cm; (b,e) E = 629 kPa, Point C = −1 cm; (c,f) E = 1243 kPa, Point C = 7 cm. Bar size: 4 μm. Images a-c are topographic images illustrating the 3D fibrous network; Images d-f are deflection images highlighting the fibril bundling and collagen D-period. The inserts are 4.5 × 4.5 μm² images taken at the same areas to highlight local fibril distribution. Arrows in (b,e) highlight the typical pore-like hollow spaces; arrows in (c,f) highlight the wounds.

2.6. Alignment index of collagen fibrils

Collagen fibril alignment was analyzed based on >15 AFM images (at 7 × 7 μm²) per patient. A 2D fast Fourier transform (2D FFT) approach was used to process 512 × 512 pixels of 8-bit gray-scale images using ImageJ software supported by an oval profile plug-in [26,32,33]. This produces a frequency plot (or power spectrum) that carries the anisotropic spatial information of the fibrils in the original AFM image. Pixel intensities were summed along the radius of each angle to produce a 2D FFT alignment plot for quantitative analysis of fibrils oriented in 0°−360° The alignment index was defined as the ratio of the highest and the lowest peak values in a FFT plot.

2.7. RT-qPCR analysis

Vaginal fibroblasts were extracted from the harvested biopsy following an established protocol [7]. Total RNA was extracted from fibroblasts using a PureLink RNA Mini Kit (Ambion, Grand Island, NY). Reverse-transcription was carried out using a SuperScript^® III kit (Invitrogen, Carlsbad, CA). RT-qPCR was performed using an ABI Prism 700 (Applied Biosystem, Foster City, CA) with TaqMan^® Gene Expression Assays (Life Technology, Madison, WI). Primers against COLI (Hs00164404_m1) and COLIII (Hs00164103_m1) were used in this study. GAPDH (Hs99999905_m1) was used as an endogenous reference. Data analysis was performed using the 2^−ΔΔCT method for relative quantification based on three replicates.

2.8. Data analysis

Linear regression analyses of the correlations between patients’ POP-Q scores and the microscopic features of collagen in the corresponding connective tissues were carried out to identify the factors associated with POP severity. Additionally, a correlation heatmap was generated based on the Pearson correlation coefficients between different variables in the patients’ data set. A correlation coefficient greater than 0.7 or lower than −0.7 was considered a strong positive or negative correlation.

3. Results

3.1. Multi-scale features of collagen in vaginal wall tissues

The nanoscopic to microscopic structures, mechanical strength, and biochemical composition of collagen in tissues of 17 subjects were examined. As shown in Fig. 1a, Gomori trichrome staining was performed to identify collagen in vaginal wall tissue specimen. The three layers of the vaginal wall, namely, the stratified squamous epithelia, the mucosa and the muscularis, were identified in this POP sample and all other samples in the study. The mucosal layer is collagen-rich (blue), and extends deep to the muscularis layer, which contains both smooth muscle (red) and collagen. Note that a small number of elastic fibers are also present in the region, and can be readily identified from collagen fibrils due to the marked difference in fiber width. We focused on the study of collagen at the mucosa region. In this region, 2–10 μm thick striated collagen fibers were observed in immunofluorescent images (Fig. 1b). COLI (green) and COLIII (red) were found co-present in these fibers. However, the expression levels varied in different samples. On the nanoscopic scale (Figs. 1c,d), these fibers were shown as bundles of locally aligned collagen fibrils (65–130 nm thick) with characteristic D-period. The D-period as well as fibril width and alignment differed in tissues from patients with varied POP conditions. Young’s modulus (E-value) of collagen fibrils in the tissue samples was quantitatively determined by a nano-indentation method. A high Young’s modulus implies stiff and less flexible fibrils [27–29]. Correlations between these properties and the patient’s POP-Q scores were analyzed (see details below).

3.2. Correlations between collagen fibril Young’s modulus and POP-Q measurements

The clinical data and collagen fibril stiffness, characterized by Young’s modulus, in tissues of 17 subjects are shown in Table 1. With the increase of POP stage, the collagen fibrils were stiffer as indicated by the increased Young’s modulus. No clear correlation was found between patients’ age or menopausal status and their collagen fibril Young’s modulus, consistent with our previous report [21]. However, a higher parity (>3) seemed to correlate to a higher fibril Young’s modulus. Among the large number of Stage III patients, the collagen Young’s modulus varied in a wide range. For Patients #11 and #13, the Young’s moduli were even higher than that of the Stage IV patient. It suggests the weak correlation between POP stage and collagen fibril stiffness. Note that a pessary was employed for Patient #10 to manage her POP symptoms. However, her collagen fibril Young’s modulus remained high.

Table 1.

Clinical data and collagen fibril Young’s modulus (E-value) in tissues of 17 subjects included in the study.

Patient #	POP Stage	Age	Parity	Menopause	E-valuea (kPa)
1	0	40	1	No	213 ± 4
2	0	41	0	No	188 ± 7
3	0	45	2	No	236 ± 8
4	II	62	3	Yes	337 ± 5
5	III	40	1	No	684 ± 6
6	III	44	2	No	793 ± 7
7	III	44	2	No	824 ± 30
8	III	52	3	Yes	924 ± 22
9	III	55	2	Yes	915 ± 15
10b	III	56	1	Yes	1064 ± 31
11	III	61	6	Yes	1243 ± 12
12	III	64	3	Yes	629 ± 11
13	III	66	6	Yes	1132 ± 46
14	III	73	3	Yes	1048 ± 10
15	III	77	3	Yes	1068 ± 15
16	III	80	2	Yes	579 ± 6
17	IV	66	2	Yes	1073 ± 23

^aE-values are indicated as mean ± standard error of >1000 measurements per subject.

^bFor Patient #10, pessary was employed to manage POP symptoms.

Linear regression analyses were carried out to examine correlations between collagen fibril Young’s modulus and each of the nine clinical POP-Q measurements for the 17 subjects (Fig. 2). Note that the data for Patient #10 was excluded from the regression calculation, but was shown as a red triangle in each graph. The measurements of Point C and Point D showed good correlations with the collagen fibril Young’s modulus, and the Young’s moduli were best correlated with the Point C values (R² > 0.9).

Fig. 2.

Linear correlations between collagen fibril Young’s modulus (E-value) and patients’ POP-Q measurements, including six reference points (a, b, e for Aa, Ba and C; c, d, f for Ap, Bp and D), and three landmarks (g-i, GH, PB and TVL). In each plot, the black spots refer to data for non-POP individuals; the red triangle refers to data for Patient #10; the square and the diamond indicate data for Patient #4 (Stage II) and Patient #17 (Stage IV), respectively. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

As shown in Fig. 2e, the fibril Young’s modulus of non-POP individuals (Point C = −8 cm) was less than 250 kPa. With the increase of Point C value, the severity of apical prolapse increased and the collagen fibrils were stiffer (increased Young’s modulus). For Patient #10 (red triangle), however, the Point C value was measured at −1 cm and the fibril Young’s modulus was 1064 kPa, high above the linear correlation curve. Apparently, while the pessary could physically support the local tissue and effectively hold the prolapsed organs into place, it didn’t improve tissue elasticity.

3.3. Correlations of collagen fibril Young’s modulus and point c measurement with collagen nanoscopic features

AFM imaging was applied to examine the nanoscopic structures of collagen fibrils in the context of a native tissue. Fig. 3 shows representative AFM images illustrating the progressive structural changes with increasing apical prolapse severity. In the connective tissue of a non-POP individual with a low collagen fibril Young’s modulus (Fig. 3a,d), long, thin and dense collagen fibrils oriented locally and intertwined in ropes or braids 2–3 μm wide. In the tissue of a POP patient with increased Young’s modulus (Fig. 3b,e), collagen fibrils were found to coalesce laterally, forming fibril bundles composed of well aligned fibrils. Meanwhile, pore-like hollow spaces were more frequently observed between the bundles (arrow-pointed regions in Fig. 3b,e), presumably due to local condensation of the fibrils when forming fibril bundles. For the patient at an advanced stage of prolapse (Point C = 7 cm) and with the collagen fibril Young’s modulus as high as 1243 kPa (Fig. 3c,f), approximately straight and parallel aligned fibrils were observed to span in large areas. Consistent with our previous observation, the fibrils were uneven and deflated [21]. Two wounds (arrow pointed in Fig. 3c,f) perpendicular to the direction of fibril alignment were apparent in this sample, likely resulting from rupture of the overstretched tissue. The intertwined fibril network as shown in Fig. 3a,b was largely absent in this and other images of tissues collected from patients with advanced POP.

Quantitatively, we determined the collagen fibril orientation index, fibril width and collagen D-period based on the AFM images. The orientation index of the fibrils was found to increase from a mean value of 0.03 for a non-POP individual (Point C = −8 cm) to 0.35 for a POP patient (Point C= +7 cm). Collagen D-period, the characteristic banding pattern, is a measure of the proper assembly of triple helix molecular collagen into fibrils and fibril network [34]. While the collagen D-period varied in a narrow range of 67.4 nm to 68.8 nm for non-POP subjects, it was reduced to 61.3 nm for the patient with a point C value of +7 cm and fibril Young’s modulus of 1243 kPa.

Linear regression analyses suggest that with the increase of both Young’s modulus of collagen fibrils and clinical Point C measurement, fibril alignment index increased linearly, whereas the collagen D-period decreased linearly (Fig. 4). The correlations are strong as indicated by the R² values. An exception is the data for Patient #10. While this patient’s fibril alignment index and D-period were close to the linear correlation curves with respect to Young’s modulus, they deviated significantly from the linear correlation curves with respect to the patient’s Point C value (Fig. 4a,b). On the other hand, the collagen fibrils were thin (65–75 nm) when the Young’s modulus and the Point C value were low. With the increase of E-value and Point C value, the fibrils became much thicker but the width varied in a narrow range (110–130 nm), and the correlation was less linear (Fig. 4c).

Fig. 4.

Linear regression analyses of correlations of collagen fibril alignment index (a), D-period (b), and fibril width (c) with collagen fibril Young’s modulus (blue) and patient’s Point C value (red). In each plot, the black spots refer to data for non-POP individuals; the triangle refers to data for Patient #10; the square and the diamond indicate data for Patient #4 (Stage II) and Patient #17 (Stage IV), respectively. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

3.4. Correlations of collagen fibril Young’s modulus and point C measurement with collagen composition

Fig. 5 shows the immunofluorescent images of the vaginal wall tissues stained for COLI (green) and COLIII (red) in the mucosal region. Consistent with our previous reports [21,22], COLI and COL-III largely co-existed and were coherently distributed. The images reveal that with the increase of Point C value and collagen fibril Young’s modulus, both COLI and COLIII were less brightly stained, suggesting decreased amount of fibrous collagen within the tissues.

Fig. 5.

Immunofluorescent images of collagen in the mucosal region of vaginal wall tissues stained for COLI (green) and COLIII (red). (a,d) E = 236 kPa, Point C = −8; (b,e) E = 629 kPa, Point C = −1; (c,f) E = 1243 kPa, Point C = 7. Bar size: 105 μm. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

The changes in samples of all 17 subjects were quantitatively examined by ImageJ analysis. With more severe prolapse, indicated by higher C value, and with increased fibril stiffness, protein content of both COLI and COLIII was reduced. Importantly, the COLI/COLIII ratio increased from a mean value of 2.0 in non-POP biopsies, to 2.3 in a mild prolapse (C = −1 cm), and to 3.4 in a severe case (C = +7 cm). Due to the distinctive function of COLI and COLIII in connective tissues, a higher COLI/COLIII ratio is associated with a stiffer fibrous matrix [21,35], and consequently less stretchable vaginal wall tissue. As shown in Fig. 6, the amounts of COLI and COLIII and the COLI/COLIII ratio linearly correlated with Point C value and fibril stiffness. The combined results suggest a progressive change of collagen composition in the tissues of prolapse patients. Note that the data for Patient #10 fitted well in the correlation curves with respect to fibril stiffness, but not to the patient’s Point C value.

Fig. 6.

Linear regression analyses of correlations of COLI amount (a), COLIII amount (b) and COLI/COLIII ratio (c) with collagen fibril Young’s modulus (blue) and patient’s Point C value (red). In each plot, the black spots refer to data for non-POP individuals; the triangle refers to data for Patient #10; the square and the diamond indicate data for Patient #4 (Stage II) and Patient #17 (Stage IV), respectively. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Fig. 7 illustrates the correlation matrix based on the Pearson correlation coefficients between different properties of collagen in the patients’ data set. In addition to the properties stated above, gene expressions of COLI and COLIII in fibroblasts extracted from patients’ tissues were also included to examine any malfunction of the collagen-producing cells. However, no strong correlation was found between the collagen gene expression level in fibroblasts and the properties of collagen protein in tissues (|R| < 0.7). The strongest correlations were found for collagen fibril Young’s modulus with fibril alignment index, D-period, and the total amount of COLI and COLIII (|R| ≥0.94). This implies that fibril stiffness is associated with the hierarchical assembly of molecular collagen into its 3D architecture. The fibril alignment index was also strongly correlated with the amount of COLI and COLIII, COLI/COLIII ratio, fibril width, and D-period (|R| ≥0.91). Thick, rigid fibrils were less flexible and aligned better. Strong associations of collagen D-period with COLI/COLIII ratio and the amount of COLI and COLIII were also found (|R| ≥0.91), suggesting that the biochemical composition of collagen plays a role in the assembly of collagen.

Fig. 7.

Correlation matrix for collagen properties. Negatively and positively related properties are indicated in blue and red, respectively. The darker the color, the stronger the correlation. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

4. Discussion

POP is frequently associated with connective tissue disorders. Abnormal amount, assembly, degradation and aging of collagen undermine connective tissue functions [12,36]. Collagen fibrils are essential elements of the connective tissue network. Their strength and stiffness play a critical role in tissue biomechanics. In examining the change of collagen fibril Young’s modulus of 3 non-POP individuals and 14 POP patients, the Young’s modulus was found to vary in a large range for tissues from patients in Stage III, and its correlation with traditional POP stage is weak. The latter is not conclusive as the variance in Stage II and Stage IV groups was not adequately examined due to limitation of the number of recruited patients at Stage II and Stage IV (one each) in this study. A traditional POP stage is in the scale of 0 to IV evaluated on the basis of the most distal portion of prolapse no matter the prolapse is anterior, posterior or apical. In this study, the Young’s modulus most strongly correlated with the value of Point C in the POP-Q system, a measure of the most distal edge of the cervix, that is directly relevant to apical prolapse (see Fig. S1). The strong correlation suggests that the collagen fibril stiffness can be an indicator of the level of apical prolapse. Point D is also relevant to apical prolapse. Its weaker correlation with collagen fibril Young’s modulus is likely due to the relatively inaccurate Point D measurement. Among other POP-Q measurements, the landmarks of GH, TVL and PB are important topographical measures, but they do not illustrate the extent of prolapse. Points Aa and Ba measure descent of the anterior vaginal wall, relevant to anterior prolapse; Points Ap and Bp measure descent of the posterior vaginal wall, relevant to posterior prolapse. Since the tissue biopsies were collected at the posterior fornix, which is directly affected by apical prolapse, the result also implies that the measured collagen fibril stiffness was characteristic of the local tissue.

Analyses of the biochemical composition and the biophysical characteristics of collagen in these patients’ tissues revealed that these properties correlated with the Point C measurement and the collagen fibril stiffness in a similar manner. These linear correlations were quite strong, either positively or negatively. Strong correlation of these properties with other POP-Q measurements was not found, suggesting that (1) the measured tissue properties are closely associated with collagen fibril stiffness and (2) they are characteristic of apical prolapse only, but less relevant to other types of prolapse. The properties of vaginal wall tissues of the same patient collected at other sites may differ. Examination of vaginal wall tissues harvested at multiple sites and evaluating their properties in correlation with various POP-Q measurements could provide insightful information and a global assessment of collagen in tissues of POP patients.

The linear regression curves (Figs. 4 and 6) imply that the prolapse progression is synchronized with the deterioration of the collagen matrix. Vaginal wall tissues of POP patients contained less collagen with an increased COLI/COLIII ratio. Additionally, the assembly of collagen fibrils was defective, evidenced by the altered collagen D-period and uneven fibril width; the fibrils were mechanically exhausted due to load bearing, supported by the increased fibril stiffness and alignment. Normally, fibroblasts in connective tissues are capable of remodeling the tissue matrix in response to alterations in the local environment. Aligned and stiff collagen fibrils can stimulate fibroblasts to increase collagen synthesis to replenish collagen for tissue repair. However, fibroblasts of POP patients have poor contractility [7,37,38], rendering them less responsive to physical cues as corroborated by the poor correlation of gene-level and protein-level collagen expression (Fig. 7). MMPs are upregulated in fibroblasts of POP patients [7,37,38]. We found this as well in our analysis of MMP expression in fibroblasts extracted from tissue biopsy (Fig. S2). Presumably, increased MMP expression promoted collagen degradation and consequently, prevented the formation and proper deposition of young fibrils from newly synthesized collagen [21,22]. As a result, aged and exhausted collagen fibrils were dominant in the tissue of POP patients.

Clinical observations suggest that descending Point C (increase of Point C value) would increase the pulling force on the fornix. In response to the directional pulling force arisen from apical prolapse, these fibrils were overly stretched, became stiff, and appeared straight and highly aligned (Fig. 3). The slow turnover of collagen likely led to a looser and more fragile collagen matrix and therefore, a more severe prolapse.

Our previous work has shown that COLIII fibrils were more flexible and mechanically weaker than COLI fibrils, hence COLIII degradation was faster than COLI [21,28]. Consistently, COLI/COLIII ratio in POP patients increased with prolapse severity. A similar observation was that increased MMP expression level in diabetic patients was accompanied by increased COLI/COLIII ratio in the collagen matrix of chronic wounds, leading to delayed tissue granulation and consequently, delayed overall wound healing [25]. COLIII regulates the lateral assembly of collagen fibrils, limiting fibril diameter due to the slower process of assembling COLIII procollagen compared to COLI procollagen in the extracellular space prior to the assembly of triple-helical collagen molecules into fibrils [21,39–42]. The higher the COLI/COLIII ratio, the thicker the hybrid fibril. This is consistent with our observation that the collagen fibrils are thicker at an increased COLI/COLIII ratio in tissues of POP patients. Due to the difference in mechanical strength of COLI and COLIII fibrils, the increased COLI/COLIII ratio is also accountable for the increased stiffness of the collagen fibrils.

Noticeably, we observed a linear decrease of collagen D-period with POP severity. The changes were small yet significant, as evidenced in our previous work [21,22,27,29]. Similar changes in collagen D-period due to strain [34,43,44], dehydration [45] and bone diseases [46,47] have been reported. The characteristic banding pattern of ∼ 67 nm D-period signifies the proper assembly of tropocollagen molecules into fibrils and fibril network to provide mechanical strength. A change in D-period is ascribed to the combination of molecular elongation/shrinkage, an increase/decrease in the gap region between longitudinally adjoining molecules along the fibril axis, and the relative slippage of laterally adjoining molecules along the fibril axis [48]. These changes are associated with elastic energy storage, transmission and dissipation [44,49,50].

In normal tissues, the D-period of collagen fibrils increases in response to a strain and recovers upon removal of the strain in support of the pelvic organs. In POP patients, however, the D-period decreased in overly stretched collagen fibrils, suggesting the fibrils had become stiff and likely involved defects in collagen crosslinks that impair their biomechanical function. It is known that lysyl oxidase (LOX) and LOX-like enzymes enable the inter-polypeptide chain crosslinks for adjacent collagen fibrils and provide mechanical integrity of the extracellular matrix. Their decreased expression in POP patients and its impacts on collagen and elastic fiber crosslinking have been reported [51–53]. On the other hand, collagen is naturally glycosylated, having sugar molecules covalently bonded to its lysine and hydroxylysine residues. Glycation can induce the formation of advanced glycation end products (AGEs), which consist of crosslinks between collagen molecules. The AGE-mediated intermolecular crosslinks can induce fibril stiffening due to loss of enzymatic crosslinks, hence, interruption of collagen molecular arrangement and the D-period [54]. Our preliminary data also showed that in vitro reconstitution of collagen extracted from POP tissues resulted in collagen fibrils exhibiting longer than normal D-period, supporting the presence of crosslinking defects in POP collagen fibrils. It is also possible that the artifact in D-period was caused by a gene mutation. A 2.5 nm change of collagen D-period found in osteoporosis was ascribed to a COLI mutation [46,47] which caused conformational changes in the triple helix and consequently in collagen fibril assembly [55]. Further study is needed to clarify the cause of the variation of tissue architecture building blocks that impact tissue mechanics.

In addition to collagen, elastin and SMCs are the primary load-bearing constituents of vaginal wall connective tissues, whose mechanical behavior is predicated on the structure, composition and organization of these constituents, as well as the effectiveness of their interdigitated network. We reported previously that, in POP vaginal wall tissues, loose fibrous collagen segregated from SMCs lead to deterioration of matrix integrity [21,22]. The abnormal extracellular environment can invoke reduced cellular capacity of SMCs to produce elastin and maintain the cross-linked elastic matrix of the vagina and pelvic floor, and hence, result in reduced mechanical strength of the tissue [52].

POP is associated with multiple factors. Multiparity, old age, overweight, physical trauma, chronic straining and obstructive lung disease are important risk factors for POP [56–58]. Hypo-estrogenic state in post-menopausal women can also impinge upon the biomechanical state of the connective tissues [59,60]. Additional factors, such as smoking, diabetes, race, COVID status, may also play some roles. Inclusion of these parameters for patient recruitment and characterization of tissues from a larger number of patients will empower data analysis to evaluate their individual and combined effects.

Severe stages of POP often require reconstructive pelvic surgery by employing permanent or degradable meshes made from materials of synthetic or biological origin [61,62]. In less severe apical prolapse cases, a pessary is applied to manage the discomfort of uterine prolapse. As shown in this study, while a pessary could physically relieve the stress of the local tissue and lift up the prolapsed organs, it didn’t rectify the aberrant biophysical and biochemical properties of collagen for tissue regeneration. Availability of more management options and affordable treatments are highly desirable. Fibroblasts are the main source of collagen. Collagen deficiency and abnormal upregulation of MMPs in POP fibroblasts have been reported [7–10]. It inspired our recent effort to restore the biological function of POP patients’ fibroblasts via matrix mediated electrical stimulation [26]. The transformed cells can be resupplied to patients for reviving the tissue function. The use of autologous vaginal wall fibroblasts to treat vaginal wall connective tissue disorder is safe, simple and inexpensive, and is expected to be an economically favorable and clinically relevant alternative to reconstructive surgery or stem cell therapy.

5. Conclusions

Taken together, collagen structure and composition in tissue matrix of POP patients are compromised. While it remains unclear if a collagen disorder is the cause or consequence of the prolapse [61,63], these properties were found to change progressively with the stiffness of collagen fibrils in the tissue and the severity of apical prolapse. Analysis of a large set of data, derived from both clinical measurements and lab-based tissue analysis, will potentially allow the use of combined POP-Q values, non-invasively measured during a patient’s clinical visit, to diagnose the local abnormity of collagen in a prolapse patient, hence forming the basis of selecting a personalized therapeutic treatment and identifying novel targets for regenerative therapies. Knowledge derived from this study has laid the groundwork for us and others to design functional prosthetic materials for use in prolapse surgery and to develop personalized therapies.

Statement of significance.

Abnormal collagen metabolism and deposition are known to associate with connective tissue disorders, such as pelvic organ prolapse. Quantitative correlation of the biochemical and biophysical characteristics of collagen in a prolapse patient’s tissue with the clinical diagnostic measurements is unexplored and unestablished. This study fills the knowledge gap between clinical prolapse quantification and the individual’s cellular and molecular disorders leading to connective tissue failure, thus, provides the basis for clinicians to employ personalized treatment that can best manage the patient’s condition and to alert pre-symptomatic patients for early management to avoid unwanted surgery.