Chemokines and Their Association With Symptom Severity in Women With Endometriosis

The study was performed in accordance with ethical principles that had their origin in the Declaration of Helsinki and was approved by both the Regional Ethical Review Board of Umeå Sweden on the 13th of October 2021 (Dnr 2021‐05236) and the Provincial Health Directorate, Kanuni Sultan Suleyman Training and Research Hospital Ethics Committee, Istanbul, Turkey (Dnr. 2021.11.281).

A total of 45 healthy controls and 90 patients were screened for the study. Of these, 62 participants were excluded because they had been diagnosed with another concurrent condition associated with pain (e.g., fibromyalgia or ankylosing spondylitis, n = 35), had been taking NSAIDs at the time of blood sampling ( n = 20), or they were not in the secretory phase (when menstrual phase was confirmed histologically from the endometrial biopsy, n = 7). Ultimately, the study comprised 22 women in the control group and 51 women in the patient group. The baseline characteristics of the participants in each group are shown in Table  1 . There were no differences in age or BMI between the two groups. Most patients ( n = 48, 94%) were diagnosed with ovarian endometrioma; the other patients had peritoneal endometriosis, one had been diagnosed with deep infiltrating endometriosis. The diagnosis of endometriosis was confirmed by surgery in 49 patients (96%) and the others were diagnosed solely by ultrasound (ovarian endometrioma). Demographic characteristics of patients with endometriosis and healthy controls. Significance testing was performed using the Mann‒Whitney U test for continuous variables and the χ 2 test for dichotomous variables. Interquartile range. Body mass index. Visual analog scale. 30‐item Endometriosis Health Profile. Twenty‐seven patients (53%) were receiving hormonal treatment ( n = 27). To determine the influence of hormonal treatment on the level and distribution of chemokines, we divided the patient group into hormone‐treated (HT) and non‐hormone‐treated (NHT) subgroups. The treatments used in the HT group were dienogest ( n = 14, 51.9%), combined oral contraceptives ( n = 5, 18.5%), and dydrogesterone ( n = 8, 29.6%), with a median treatment duration of 10 (8.0‒14.0) months. The severity of endometriosis‐related pain symptoms was assessed in the patient group using the VAS and EHP‐30. We found no significant differences in VAS or any dimension of the EHP‐30 questionnaire between the HT and NHT subgroups. To investigate the differences in chemokines (CCL2, CXCL8, and CXCL1) between the control and patient groups, we compared the serum concentrations and endometrial expression between the two groups. The serum levels of CCL2 and CXCL1 were significantly lower in the patient group than in the control group (both, p < 0.05), and the endometrial expression of CXCL8 was significantly decreased in the patient group ( p < 0.05). The serum concentrations and endometrial expression of chemokines in both groups are shown in Table  2 . Immunohistochemistry of endometrial CCL2, CXCL8, and CXCL1 in patients and controls are shown in Figure  1 . Chemokine serum levels (pg/mL) and endometrial expression (HSCORE a ) in women with endometriosis and healthy controls. Significance testing was performed using the Mann‒Whitney U test. HSCORE, immunohistochemical intensity score. Serum and endometrial analysis based on 50 available samples from 51 included patients; one patient was missing serum analysis and one patient was missing an endometrial sample. IQR, interquartile range. Endometrial chemokines (arrow) assessed by immunohistochemistry, in healthy controls; (A) CCL2, (B) CXCL8, (C) CXCL1, and in women with endometriosis; (D) CCL2, (E) CXCL8, (F) CXCL1. To investigate whether the serum level of each chemokine was correlated with its endometrial expression, we determined the correlation between the serum level and endometrial staining intensity (HSCORE). However, we found no correlations between the serum level and endometrial expression of any of the chemokines in either group. We assessed whether the different measures of symptom severity (VAS and EHP‐30) were correlated in the patient group. However, we found that the VAS was not correlated with the five domains of health‐related quality of life (pain, control and powerlessness, emotional well‐being, social support, and self‐image). We performed regression analyses to determine whether the serum levels or endometrial expression of the three chemokines (CCL2, CXCL8, and CXCL1) were associated with the severity of symptoms reported by the patients (VAS and the pain domain of the EHP‐30 questionnaire). Increased endometrial expression of both CXCL8 and CXCL1 was associated with higher VAS scores ( p < 0.05 and p < 0.001, respectively). Additionally, increased serum levels of all three chemokines (CCL2, CXCL8, and CXCL1) were associated with increased scores of the EHP‐30 pain domain ( p < 0.05 , p < 0.05, and p < 0.001, respectively; Table  3 ). Association between serum levels (ng/mL) and endometrial expression (HSCORE † ) of chemokines with symptom severity (VAS and pain dimension of the EHP30 b ) in women with endometriosis. Linear regression analysis was performed with adjustment for age, body mass index, and ongoing hormonal treatment as confounding parameters. VAS β c (95% CI d ), p EHP30, Pain β c (95% CI d ), p 0.000 (−0.003–0.003) Ns 0.058 (0.015–0.100) < 0.05 −0.002 (−0.011–0.007) Ns 0.159 (0.030–0.288) <0.05 0.000 (−0.001–0.000) Ns 0.024 (0.017–0.032) <0.001 VAS β c (95% CI d ), p EHP30, Pain β c (95% CI d ), p 0.005 (0.000–0.010) Ns −0.026 (−0.104–0.051) Ns 0.009 (0.004–0.015) <0.05 −0.092 (−0.185–0.002) Ns 0.009 (0.005–0.013) <0.001 −0.46 (−0.121–0.028) Ns Immunohistochemical intensity score. 30‐item Endometriosis Health Profile. Regression coefficient. Confidence interval. To investigate how the chemokines (CCL2, CXCL8, and CXCL1) differed between the HT and NHT subgroups, we compared the serum level and endometrial expression of these chemokines between these subgroups. Except for an increased serum level of CCL2 in the HT group ( p < 0.05), there were no significant differences in the serum levels and endometrial expression of the chemokines between the two subgroups (Table  4 ). Demographic characteristics, serum chemokine levels and endometrial chemokine expression in women with endometriosis with and without hormonal treatment. Significance testing was performed with Mann–Whitney U test. No Hormone treatment Median (IQR d ) ( n = 24) Hormone treatment Median (IQR d ) ( n = 27) Note: Demographic characteristics, serum chemokine levels and endometrial chemokine expression in women with endometriosis with and without hormonal treatment. Significance testing was performed with Mann–Whitney U test. HSCORE, immunohistochemical intensity score. 30‐item Endometriosis Health Profile. Visual analog scale. Interquartile range. The association between CXCL1 and symptom severity remained in the HT and NHT subgroups. In the HT subgroup, increased endometrial expression of CXCL1 was associated with higher VAS scores ( β = 0.008, 95% CI: 0.001–0.015 , p < 0.05) and increased serum levels of CXCL1 were associated with higher EHP‐30 pain scores ( β = 0.028, 95% CI: 0.009–0.047 , p < 0.05). Likewise, in the NHT subgroup, increased endometrial expression of CXCL1 was associated with higher VAS scores ( β = 0.011, 95% CI: 0.007–0.016 , p < 0.001) and increased serum levels of CXCL1 were associated with higher EHP‐30 pain scores ( β = 0.022, 95% CI: 0.014–0.030 , p < 0.001) (Figure  2 ). By comparison, neither the serum level, nor the endometrial expression of CCL2 was associated with any measure of symptom severity in the HT and NHT subgroups. The association between serum CXCL8 and the EHP‐30 pain score remained in the HT subgroup ( β = 0.207, 95% CI: 0.017–0.396 , p < 0.05) but not in the NHT subgroup. The association between endometrial CXCL8 expression and VAS remained in the NHT subgroup ( β = 0.010, 95% CI: 0.003–0.016 , p < 0.05) but not in the HT subgroup. The association between CXCL1, CCL2, and CXCL8 and symptom severity (assessed by the pain dimension of the EHP‐30 and VAS) in women with endometriosis, with and without hormone treatment.

All participants were recruited from the outpatient clinic at Kanuni Sultan Suleyman Training and Research Hospital in Istanbul, Turkey. The inclusion criteria were women aged 18‒45 years with suspected endometriosis, confirmed by either laparoscopy or ultrasound (patient group). The control group consisted of healthy women undergoing sterilization by tubal ligation. The exclusion criteria were ongoing pregnancy, diagnosis of a concurrent condition associated with inflammation or pelvic pain, and current regular use of non‐steroidal anti‐inflammatory drugs (NSAIDs). The study was designed as a cross‐sectional study. Women were recruited after they had received oral and written information about the study and signed the consent form. All participants were examined in the secretory phase of the menstrual cycle, assessed during the clinical review and physical examination, and later confirmed by histological examination of the endometrium. Patients with ongoing hormone treatment, all of whom were amenorrhoeic, were examined without considering the menstrual cycle phase. All participants underwent general physical examinations, which included gynecological and ultrasound examinations. After the assessment of symptom severity, an endometrial biopsy was performed for histopathological confirmation of a secretory endometrium and for immunohistochemical assessment of endometrial chemokine expression. A blood sample (5 mL) was drawn to measure serum chemokine levels. All patients assessed their current level of pain using a visual analog scale (VAS) and completed the Endometriosis Health Profile 30 (EHP‐30) questionnaire to assess the impact of symptom severity on health‐related quality of life. The VAS is one of the most well‐known methods for examining a patient's pain severity. The patient reported their present overall pain by placing a marker or cross on a 10 cm line, which ranged from 0 = no pain to 10 = worst imaginable pain. EHP‐30 is a questionnaire that was developed to measure the extent to which endometriosis has a negative impact on the patient's quality of life. It quantifies symptom severity in endometriosis by assessing how endometriosis‐associated symptoms have affected the patient's quality of life for the past 4 weeks. It comprises the following five dimensions of health‐related quality of life: pain, control and powerlessness, emotional well‐being, social support, and self‐image. Each dimension is scored separately from 0 to 100, where 100 indicates the worst possible health‐related quality of life [ 14 ]. The questionnaire is commonly used to evaluate the effects of medical and surgical therapies for endometriosis [ 15 ]. Five milliliters of blood were collected from each participant. The samples were centrifuged at 3000 × g for 10 min to separate the serum component. Serum samples were stored at −80°C until the biochemical assays were performed. The serum concentrations of CCL2 (MCP‐1, SunRedBio, China), CXCL8 (IL‐8, SunRedBio), and CXCL1 (GRO‐α, SunRedBio) were quantified using enzyme‐linked immunosorbent assay kits following the manufacturer's instructions. The samples were measured by spectrophotometry (Synergy HTX, BioTek Instruments, Winooski, VT, USA) at a wavelength of 450 nm. The levels of CCL2, CXCL8, and CXCL1 in the samples were determined by comparing their optical densities with established standard curves and are expressed as picograms per milliliter (pg/mL). Formalin‐fixed, paraffin‐embedded tissues were cut into 3‐µm‐thick sections, which were transferred to a positively charged slide. Deparaffinized, dehydrated sections were placed in distilled water. For the pre‐enzyme procedure, EnVision FLEX target retrieval solution high pH was added to the pretreatment device (Dako PT Link, CA, USA). The pre‐heat temperature was 65°C and the epitope retrieval temperature was 97°C. The tissues were kept in the device for 20 min, cooled to 65°C, immediately transferred to a jar containing Tris‐buffered saline buffer (ScyTek, West Logan, UT, USA), and washed with the solution twice. Then, 3% H 2 O 2 solution was applied for 10 min at room temperature. The samples were incubated with rabbit anti‐human CCL2 (GeneTex, Irvine, CA, USA), rabbit anti‐human CXCL8 (GeneTex), and rabbit anti‐human CXCL1 (GeneTex) at a dilution of 1/200 for 2 h. SensiTek anti‐polyvalent biotinylated antibodies (ScyTek) were applied for 20 min followed by SensiTek horseradish peroxidase (ScyTek) for 20 min. Next, slides were immersed in 3,3′‐diaminobenzidine chromogen substrate solution for 10 min at room temperature, and washed in distilled water. Slides were counterstained by applying hematoxylin stain for 1 min. The slides were washed in distilled water and dried. Lamellas adhered to the slides via a mounting medium. Immunohistochemical staining of CCL2, CXCL8, and CXCL1 was evaluated by a single examiner who was blinded to the participant group and degree of symptom severity. The evaluation was performed in a semiquantitative fashion using a grading system from 0 (no staining) to 3 (most intense staining). For each slide, 10 areas were evaluated and the HSCORE value was derived by summing the percentages of stained cells at each intensity level, multiplied by the intensity score [HSCORE = ∑P i ( i + 1)], where i represent the intensity score and P i is the corresponding percentage of stained cells [ 16 ]. All statistical analyses were performed using SPSS version 28 (SPSS, IBM Corp., Armonk, NY, USA). Continuous variables are presented as medians (interquartile range). Categorical variables are presented as frequency counts and percentages. The Mann‒Whitney U test was used to compare continuous variables between groups and the χ 2 test was used to compare dichotomous variables between groups. Correlations were assessed using Spearman's rank correlation coefficient. Linear regression was used to assess associations between study parameters and results are presented as β coefficients with 95% confidence intervals (CI). Age, body mass index (BMI), and ongoing hormonal treatment were included in the regression analyses as possible confounders. p values of < 0.05 were considered statistically significant. Sample size: As the basis for the sample size calculation, we considered the results of an earlier study of patients with an orofacial pain condition that showed a positive correlation between CXCL8 and pain [ 17 ]. In our study, using a two‐sided test at a significance of 5% ( α = 0.05) and 80% power ( β = 0.2), the required sample size was approximately 22 women in each group.

We showed that increased serum levels and endometrial expression of chemokines are associated with increased symptom severity in women with endometriosis. Our findings indicate that the chemokines CCL2, CXCL8, and CXCL1 are associated with poorer health related quality of life in women with endometriosis. Moreover, increased endometrial expression of CXCL8 and CXCL1 was linked to more severe pain reported by the same patients. Among these chemokines, CXCL1 demonstrated the most consistent association, as both elevated serum levels and heightened endometrial expression correlated with more severe symptoms, irrespective of hormonal treatment status. However, given the cross‐sectional nature of the study, these findings should be interpreted as correlations rather than evidence of causation. It remains unclear whether elevated chemokine levels cause or contribute to pain severity, or whether they simply reflect a shared underlying inflammatory process. Further longitudinal or mechanistic studies are needed to clarify the causal relationship between these chemokines and endometriosis‐associated pain.

In this cross‐sectional study of women with endometriosis, we found that the serum levels of both CCL2 and CXCL1 were significantly lower among women with endometriosis than in healthy controls. Additionally, we found that endometrial expression of CXCL8 was reduced in women with endometriosis. This was surprising because our hypothesis was that women with endometriosis would have increased chemokine levels. However, all the chemokines examined in this study were associated with more severe symptoms in the patients. Increased serum chemokine levels were associated with higher scores for the pain dimension of the EHP‐30 questionnaire, and increased endometrial expression of CXCL8 and CXCL1 was associated with higher VAS scores for pain. However, when the patient group was divided into the HT and NHT subgroups according to use of hormonal treatment, CXCL1 was the only chemokine that consistently was associated with increased symptom severity (VAS and pain dimension of EHP‐30), both when we considered serum levels and endometrial expression, in both subgroups. Earlier studies focusing on other painful conditions, such as degenerative disc disease, revealed that the severity of pain symptoms was correlated with both the local expression and serum levels of chemokines, including CCL2 and CXCL8 [ 18 , 19 ]. Although few studies have explored the role of chemokines in the mechanisms of pain in women with endometriosis, one study showed that polymorphisms of CXCL8, which are associated with increased CXCL8 expression, were associated with an increased risk of chronic pelvic pain in women with endometriosis [ 20 ]. We used VAS and the EHP‐30 pain dimension to assess symptom severity. Apart from differing in terms of how they measure symptom severity (pain versus the impact of pain on health‐related quality of life), the two instruments differ in the timescale being measured. The VAS represents the patient's pain at the moment the patient records the score, whereas the EHP‐30 covers a longer timescale, by recording the severity of symptoms over the past 4 weeks. Higher endometrial expression of chemokines (CXCL8 and CXCL1) was associated with higher VAS scores but not with higher EHP‐30 pain scores, and higher serum levels of chemokines (CCL2, CXCL8, and CXCL1) were associated with higher EHP‐30 pain scores but not with higher VAS scores. Because we only included patients in the secretory phase of the menstrual cycle, we speculate that this difference could be explained by a possible increase in both pain (measured by VAS) and endometrial expression of chemokines during the secretory phase, whereas the serum chemokine levels may remain more stable during the menstrual cycle and show stronger correlations with symptom severity assessed by a tool measuring symptoms over a longer period of time, like the EHP‐30. Some studies have examined the fluctuations in serum and endometrial chemokines during the menstrual cycle that could support this explanation. For example, Ulukus et al. reported that the endometrial expression of both CCL2 and CXCL8 fluctuates during the menstrual cycle, and that their expression increases during the secretory phase [ 21 ]. Regarding serum chemokine levels, at least CCL2 appears to remain stable during the menstrual cycle according to Dahm‐Kähler et al. [ 22 ]. Although the evidence is scarce, we believe that this is the most likely explanation for why increased endometrial expression of CXCL8 and CXCL1 was associated with higher VAS scores but not EHP‐30 pain scores, and why increased serum levels of CCL2, CXCL8, and CXCL1 were associated with higher EHP‐30 pain scores but not VAS scores. Over the last few decades, there has been great interest in chemokines and their role as possible biomarkers for endometriosis [ 23 ]. Earlier studies of CCL2 reported contradictory results, but most studies found either no difference or increased serum and peritoneal concentrations of CCL2 in women with endometriosis, except for one study that found lower levels in the peritoneal fluid of women with endometriosis [ 5 , 13 ]. Studies of CXCL1 reported increased levels in the peritoneal fluid of women with endometriosis, but the only study that compared serum levels of CXCL1 found no difference between women with endometriosis and controls [ 24 , 25 , 26 ]. Surprisingly, when we compared serum chemokine levels between women with endometriosis and healthy controls, none of the chemokines were elevated in the endometriosis group. Contrary to our expectations, CCL2 and CXCL1 levels were significantly lower in women with endometriosis compared to healthy controls. This finding may be explained by a combination of biological and methodological factors. Chronic chemokine elevation is known to downregulate receptor expression, which may subsequently reduce chemokine production [ 27 ]. It can be speculated that such receptor downregulation negatively impacts chemokine synthesis. Hormonal treatments commonly used by patients may further suppress immune activity and mask inflammatory signals. Moreover, the menstrual cycle phase at the time of sampling could influence chemokine levels, given their cycle‐dependent expression [ 28 , 29 ]. Endometriosis itself is characterized by systemic and localized immune suppression, including altered NK‐cell function and macrophage populations, which may blunt chemokine responses [ 30 ]. Our finding of decreased endometrial expression of CXCL8 among women with endometriosis differs from previous findings, which revealed increased endometrial expression of CXCL8 in women with endometriosis [ 21 ]. This discrepancy could be explained by the fact that the participants in our study were examined during the secretory phase of the menstrual cycle. In contrast, the article cited above reported increased endometrial expression of CXCL8 in women with endometriosis, but only during the proliferative phase of the menstrual cycle. Another important difference lies in our inclusion of patients who were receiving hormonal treatment; these patients exhibited the lowest endometrial expression of CXCL8. When these patients were excluded, we found no difference in endometrial CXCL8 expression between the patients and controls. Progesterone has immunosuppressive properties and suppresses chemokine production, but women with endometriosis exhibit progesterone receptor alterations and progesterone resistance. It is also unclear whether hormonal treatment and synthetic progestins in women with endometriosis affect chemokines [ 31 , 32 ]. When we compared the serum levels and endometrial expression of chemokines between the HT and NHT subgroups, we found no significant differences except for increased serum level of CCL2 in the HT subgroup, although it is uncertain whether this was caused by the hormonal treatment. Interestingly, when the patients were subdivided based on hormonal treatment, CXCL1 was the only chemokine for which both increased endometrial expression and increased serum concentration were associated with worse symptoms, regardless of whether the patient was receiving hormonal treatment. Although this association does not necessarily mean that CXCL1 contributes to the pain experienced by women with endometriosis, a causal relationship is supported by previous findings in rodent models, which demonstrate that CXCL1 plays a crucial role in nociception, in the development of both peripheral and central sensitization [ 9 ]. Our study population mainly comprised patients with ovarian endometrioma. This is a limitation when drawing general conclusions on endometriosis because only a handful of patients with other types of endometriosis were represented in the study. Another important limitation is the lack of detailed pain assessment. Clinically distinct types of pelvic pain such as dysmenorrhea, non‐menstrual pelvic pain, and dyspareunia may arise from different underlaying pathophysiological mechanisms. These distinctions were not captured in this study, which may limit the clinical interpretation of pain‐related outcomes. Pain was assessed solely through a general VAS reducing the granularity of the findings. Additionally, pain data from the control group were unavailable, as the VAS was not completed by the women in that group. Although the controls were considered healthy and endometriosis was ruled out during tubal ligation, we cannot exclude the possibility that some may have experienced pain at the time of enrollment.

Endometriosis is a chronic inflammatory condition defined by endometrial‐like glands and stroma at extrauterine sites. The condition affects the quality of life of patients mainly because of various pain symptoms such as dysmenorrhea, dyspareunia, and chronic pelvic pain [ 1 ]. Although the mechanism underlying the pain is not fully understood, immunological factors may play an important role [ 2 , 3 ]. Chemokines are important mediators of inflammation. There are two major chemokine subfamilies defined by the position of cysteine residues, CXC and CC chemokines [ 4 ]. Both families of chemokines have been linked to the pathogenesis of endometriosis, and widely studied chemokines include CCL2, CXCL8, and CXCL1 [ 5 ]. CCL2 belongs to the CC‐family, and its main function is to recruit macrophages to sites of inflammation, although it is also thought to promote the development of neuropathic pain by acting on its receptor, CCR2 [ 6 ]. CXCL8, a member of the CXC family, is secreted by macrophages, and promotes inflammation and angiogenesis as well as nociception by acting on CXCR1 and CXCR2 receptors [ 7 , 8 ]. CXCL1 belongs to the CXC family and is secreted by macrophages. Apart from promoting inflammation, CXCL1 also promotes nociception at nerve endings and central sensitization by acting on its receptor CXCR2 [ 9 ]. Because of their role in nociception, chemokines and their receptors are promising targets for treating pain [ 10 , 11 ]. Although previous studies have investigated the associations between chemokines, fertility scores, and the disease stage in endometriosis, it is unknown whether chemokines are involved in the pain symptoms associated with endometriosis [ 12 , 13 ]. We hypothesized that increased production, serum levels, and endometrial expression of CCL2, CXCL8, and CXCL1 would be associated with worse pain symptoms in women with endometriosis. In this study, we tested the hypothesis in the following ways. Investigated the associations between the serum concentrations and endometrial expression of the chemokines CCL2, CXCL8, and CXCL1 and the severity of pain symptoms in women with endometriosis. Evaluated whether the serum levels and endometrial expression of these chemokines differed between women with endometriosis and healthy controls. Determined whether the serum levels and endometrial expression of these chemokines differed between patients with endometriosis treated with or without hormonal therapies. Investigated the associations between the serum concentrations and endometrial expression of the chemokines CCL2, CXCL8, and CXCL1 and the severity of pain symptoms in women with endometriosis. Evaluated whether the serum levels and endometrial expression of these chemokines differed between women with endometriosis and healthy controls. Determined whether the serum levels and endometrial expression of these chemokines differed between patients with endometriosis treated with or without hormonal therapies.

The authors declare no conflicts of interest.