A comparative analysis of fixatives for improved histological integrity in the human ovary

Dear Editor,

The advent of ovarian tissue cryopreservation (OTC) for fertility preservation has allowed for the evaluation of normal human ovarian tissue. Fixation is a critical initial step in preserving the morphological integrity of specimens to prevent autolysis and degradation, thus facilitating microscopic evaluation for pathologic and research analyses [1, 2]. Extensive research has shown that the method used to handle and process tissues can significantly influence the preservation of tissue structure, proteins, and nucleic acids [3–5].

The number of human follicles, structures comprised of the oocyte and surrounding granulosa cells, is fixed at birth and thereafter declines through a process of activation and development through different stages of folliculogenesis and atresia [6]. Accurately assessing follicle development through histology is important for understanding this process. However, tissue processing and fixation may interfere with the proper assessment of follicle morphology. Although the effect of fixative type on follicle appearance has been studied in animal models, there are minimal published reports evaluating fixatives in human ovarian tissue [7]. Neutral Buffered Formalin (NBF), though commonly used in pathology departments, is known to cause shrinkage in many tissue types [8]. One study using human ovarian tissue found that NBF with 5% acetic acid (Form-Acetic) exhibited greater preservation of tissue integrity when compared to NBF, was comparable to Bouin’s fixative, and was compatible with common histological staining methods [8]. Here, we seek to evaluate the effects of different fixatives on follicle appearance through a comparative analysis of human ovarian tissue treated with three different fixatives: NBF, modified Davidson’s Fixative (mDF), and paraformaldehyde (PFA). We selected NBF as it is the standard fixative in pathology departments and used mDF and PFA due to their commercial availability. Bouin’s fixative was excluded because of significant safety concerns. As ovarian tissue is limited and immediate fixation is essential, we relied on these readily available fixatives.

The ovarian tissue was obtained by taking two consecutive “salami” slices from the central-lateral aspect of the ovary, ensuring that no more than 20% of the ovary was excised during oophorectomy for OTC [6]. Additionally, both ovaries from a 32-year-old cadaveric organ donor were obtained before stopping cardiovascular support and were cut into sections based on the NICHD ovarian nomenclature 3D description of the ovary [6]. All tissue collection was performed under IRB-approved protocols.

Tissue samples from the first slice (n = 8) were fixed in 10% NBF and embedded for routine histologic analysis at three institutions using established pathology department protocols at the site where tissue was collected. See Supplementary Material for further details.

The second slice of tissue was divided in half and fixed with mDF (n = 6) or freshly prepared 4% PFA (n = 2), while the remaining half was saved for other experiments. After overnight fixation, the mDF and PFA fixatives were replaced with 4% sucrose. Of the eight samples, seven were stored in 4% sucrose for 14 h and one was stored in 4% sucrose for 8 h. They were then stored in 70% ethanol before being embedded with paraffin. Additionally, samples from nine different regions [6] of the left ovary of the cadaveric donor were divided into two pieces and fixed with either NBF and 4% PFA. After 24 h, the tissue was replaced with 4% sucrose for 14 h and stored in 70% ethanol before being embedded with paraffin.

Six 5 μm sections were processed from each FFPE block and stained with hematoxylin and eosin (H&E). Whole-slide images (WSI) of H&E slides were scanned at 40x magnification with a Hamamatsu Nano Zoomer S60 scanner and added to the NICHD/Oncofertility ovarian tissue digital database [9] for analysis. All WSI were evaluated in a double-blind manner by at least two independent researchers. Follicles were counted and classified according to the NICHD Ovarian Nomenclature Workshop Follicle Subgroup Classification [10]. Normal preantral follicles include primordial, transitional primordial, primary, transitional primary, and secondary follicles. Follicles with abnormal features are categorized based on the affected structures. These abnormalities are classified into three categories: AMF-o, referring to follicles with vacuoles or degeneration in the oocyte; AMF-g, involving vacuoles or degeneration in the granulosa cells; and AMF-og, which includes follicles exhibiting vacuoles or degeneration in both the oocyte and granulosa cells. Among samples with these abnormalities, the vast majority of follicles were found to be primordial, transitional primordial, primary, and transitional primary [6, 9, 10].

The total percentage of abnormal follicles between the two fixatives was analyzed using a paired t-test. Additionally, a two-way ANOVA was conducted to compare the three categories of AMF across the different fixative pairs.

The percentage of abnormal follicles was significantly lower in the samples fixed with mDF compared to NBF (10.7 ± 2.2 vs. 72.1 ± 9.2; P = 0.002) (Table 1; Figure 1A). Representative sections of ovaries and individual follicles fixed with NBF and mDF are displayed in Figure 1B and C. Additionally, there was also a significantly lower percentage of abnormal follicles in the samples fixed with PFA compared to NBF (85.2 ± 2.2 vs. 95.8 ± 1.3; P = 0.021) (Table 1; Figure 1D). Representative sections of ovaries and individual follicles fixed with NBF and PFA are displayed in Figure 1E and F.

Table 1.

Demographics and total percentage of abnormal follicles in ovarian tissue fixed with mDF, PFA, or NBF

Patient	Age	Diagnosis	Fixative: NBF (% Abnormal follicles)	Fixative: mDF (% Abnormal follicles)
A	3	Neuroblastoma	85.5	5.07
B	5	RASGRP1 deficiency	84.1	17.0
C	5	HLH	32.4	8.13
D	10	Galactosemia	59.6	17.6
E	12	DOR/TS	78.3	10.0
F	21	SCD	92.9	6.25

			Fixative: NBF	Fixative: PFA
G	5	Ewing sarcoma	95.0	60.3
H	7	RMS	82.5	57.3
I	32	DCD	97.4*	91.0*

HLH, hemophagocytic lymphohistiocytosis; DOR/TS, diminished ovarian reserve/turner syndrome; SCD, sickle cell disease; RMS, cervical/uterine botryoid rhabdomyosarcoma; DCD, donation after circulatory death.

^*Patient I was a cadaver ovary sectioned into 12 different regions. Each region was counted individually; however, only nine regions were fixed with either PFA or NBF. The mean percentage of abnormal follicles in these nine regions is used for comparison

Figure 1.

(A) Percentage of abnormal follicles in ovaries fixed with NBF vs. mDF (n = 6). (B) Representative images of H&E-stained ovarian tissue fixed with NBF or mDF (Scale bar = 100 μm). (C) Representative images of follicles in ovarian tissue fixed with NBF vs. mDF from three patients (i–iii) (Scale bar = 25 μm). (D) Percentage of abnormal follicles in ovaries fixed with NBF vs. PFA (n = 11) (E) Representative images of H&E-stained ovarian tissue fixed with NBF or PFA (Scale bar = 100 μm). (F) Representative images of follicles of ovarian tissue fixed with NBF or PFA from 3 patients (iv–vi) (Scale bar = 25 μm). (G) Distribution of abnormal follicle % separated by the following abnormality types: oocyte (AMF-o), granulosa cells (AMF-g) or both (AMF-og) in ovarian tissue fixed with NBF or mDF. (H) and fixed with NBF or PFA. (I) Follicle abnormalities observed in NBF: Presence of vacuoles in the oocyte (AMF-o) (top left), evidence of oocyte shrinkage (AMF-o) (bottom left), abnormal granulosa cells due to oocyte shrinkage (AMF-g) (top right), both the oocyte and granulosa cells exhibit abnormalities (AMG-og) (bottom right) (Scale bar = 25 μm). ^*P < 0.05, ^**P < 0.01.

Although all types of abnormal follicles were represented, the majority contained abnormal oocytes (AMF-o) (Figure 1G and H). There was a significantly lower percentage of AMF-og in the samples fixed with mDF compared to NBF (P = 0.045) (Figure 1G). There was also a significant difference between the percentages of AMF-o (P = 0.005) and AMF-og (P = 0.01) in the samples fixed with PFA compared to NBF (Figure 1H). The predominant abnormal features present in samples fixed with NBF were the presence of vacuoles in the oocyte (Figure 1I, AMF-o, top left), shrunken ooplasm in oocytes (Figure 1I, AMF-o, top right), abnormal granulosa cells due to oocyte shrinkage (Fig. 1I, AMF-g, bottom left), and abnormalities in both the oocyte and the granulosa cells (Figure 1I, AMF-og, bottom right). Primordial follicles were the most affected; however, it is unclear whether this is a result of the fixatives used or the fact that our samples were from young patients, who naturally have a higher proportion of primordial follicles.

In summary, this comparative analysis of human ovarian tissue fixed with different agents highlights the significant impact of fixative choice when assessing follicle morphology and suggests that mDF may better preserve the structural integrity of ovarian follicles than traditional NBF used by pathology. Although a direct comparison in the same ovarian tissue between mDF and PFA could not be done in this study due to tissue quantity limitations, there was a higher percentage of abnormal follicles in ovarian tissue fixed with PFA relative to mDF.

We acknowledge that the distribution of follicles within the ovaries is not uniform. However, the use of tissues from adjacent sections of the same ovary reduces the possibility of atretic appearance due to age or underlying disease differences. Additionally, all three fixatives (NBF, mDF, and PFA) were not compared against each other within the same tissue sample, which hinders direct comparisons between mDF and PFA. We also acknowledge that variations in processing steps across different institutions may serve as a confounding variable. Despite these limitations, our findings suggest that the primary cause of “abnormal follicles,” particularly AMF-o due to ooplasm shrinkage, is more likely due to artifacts arising from tissue processing techniques, particularly the fixatives used, rather than overt effects from the patient’s medical condition. This data provides compelling evidence that mDF is a superior fixative compared to the commonly used NBF for preserving the morphology of preantral follicles in human ovarian tissue. Although the implementation of mDF in a clinical setting would be ideal given the scarcity of normal human ovarian tissue and the importance of optimizing the evaluation of follicle quality in ovarian tissue prior to ovarian tissue transplantation, several challenges remain. mDF is known for its toxicity, is more expensive, and the use of mDF requires multiple steps, increasing the workload for pathology departments.

Ultimately, these findings highlight the importance of not over-interpreting results obtained from tissues fixed with traditional methods and demonstrate the superiority of mDF for accurate follicle morphology assessment and the need for continued research on optimal fixation and evaluation of human ovarian tissue.

Conflict of Interest: The authors have declared that no conflict of interest exists.

Data availability

The data underlying this article will be shared on reasonable request to the corresponding author.