Optimization of ovine bone decalcification for increased cellular detail: a parametric study

Abstract

There are many published methods of decalcifying bone for paraffin histology; however, the current literature lacks details regarding the processing of ovine tissue. Ovine bone tissue presents challenges, as samples are often denser and larger than other comparative animal models, thus increasing decalcification times. Trifluoroacetic Acid (TFAA) has previously been used to decalcify ovine bone samples for histological analysis. Unfortunately, TFAA is a strong acid and often results in loss of cellular detail, especially in the connected soft tissue. This is generally manifested as over staining with eosin, and a decrease in cellular features which impacts overall histological interpretation. It is well known that leaving tissue in acid for long periods degrades cellular detail; therefore, minimizing decalcification time is critical to accurately determine cellular morphology. Six decalcification solutions (8% TFAA, 20% TFAA, 8% formic acid, 20% formic acid, Formical-4, and XLCal, and three temperatures (room temperature, 30°C, 37°C), were examined to determine their effects on cellular detail in ovine vertebrae and humeral heads. These data clearly indicate that 20% formic acid at 30°C yielded better decalcification rates (2.6 d ± 0.9 d) and cellular detail (none to mild changes) for the vertebrae samples, and 20% formic acid at RT yielded the best cellular detail (none to minimal loss) for humerus samples with a moderate amount of time (6.5 d ± 1.7). These results should establish the optimal demineralization procedures for ovine bone used in scientific studies resulting in improved cellular detail while minimizing decalcification times.

Introduction

During the processing of osseous samples for routine paraffin histology, it is necessary to remove calcium deposits through a process known as decalcification. Decalcification procedures for murine and lapin bone research is well documented [1–5]; however, current literature lacks detailed, optimized procedures used to demineralize ovine bone. Due to the differing bone size, density, and microstructure in murine bone as compared to large animal models [6], different decalcification procedures must be used for the respective species and specific bone being studied. Furthermore, the existence of soft tissue attachments (defined as tendon and spinal disc in this study) to the bone samples can create additional complications, as soft tissue is adversely affected by the decalcification procedure. Recently, an increasing amount of research has begun to focus on the attachment of tendons to bone, and the disc space within the functional spinal unit [7–10]. Optimized procedures that leave the soft tissue/bone transition unaffected are crucial for the histological analysis in these studies. The lack of a detailed ovine decalcification procedure is an issue that needs to be addressed due to the widespread use of ovine models in preclinical translational orthopedic studies [6–8,10]. These studies have become increasingly common due to the similarity in animal size, bone structure, and healing cascade in sheep as compared to humans. Therefore, a critical need exists to standardize and optimize the demineralization procedure for ovine osseous tissues. Specifically, there exist two competing interests that must be considered, which we intend to optimize: (1) total time to complete decalcification and (2) minimize an adverse impact on cellular detail within both the bone and the soft tissue attachments.

The two main types of decalcifying agents are acids and chelating agents. Trifluoroacetic acid (TFAA), a strong acid, and Formical-4 (1214, StatLab, USA) a commercial solution using formic acid, a weak acid, have been previously described as procedures for decalcifying ovine tissue [7]. Ethylenediaminetetraacetic acid (EDTA) is a chelating agent and is relatively gentle on the tissue but can take several months to fully demineralize large tissue samples.

Recent research has focused on investigation of the enthesis zone, the fibrocartilaginous transition where tendon integrates into bone, specifically the ovine infraspinatus tendon [11–13]. Due to the nature of this analysis, decalcification of the bone and soft tissue construct for paraffin sectioning is desired. However, the large amount of softer tendon creates inherent difficulties in the decalcification process. Work by our group has previously decalcified ovine humeral head/infraspinatus tendon constructs (HTCs) and functional spinal units (FSUs) with 8% TFAA at room temperature (RT) with agitation [10]. However, upon analysis of the histological sections, it was evident the TFAA decalcification protocol adversely impacts the cellular detail and overall histological quality of tendon and disc in the section. This negatively affected the ability of the pathologist to provide accurate sample interpretation. This loss of detail was manifested as diffuse over staining by eosin and an overall decrease in basophilic staining of nuclear and cytoplasmic features. In a well-stained section, nuclear and cytoplasmic features are observable due to the well-balanced color variation and contrast provided by appropriate hematoxylin and eosin staining methods. Preserving these details during the decalcification process allows the histopathologist to discern specific cell types from each other (i.e. lymphocytes from neutrophils, from macrophages, etc.). Loss of these cellular details specifically corresponds to loss of basophilic staining intensity, which impairs accurate cell-type identification, and eosin over-staining which results increasing difficulty in differentiation of eosin-stained tissue elements. Therefore, the goal of this research was to determine an optimal method of decalcification for ovine bone samples without compromising the attached tendons (HTC’s) and spinal disc (FSU’s).

Materials and methods

Ovine humeral head/infraspinatus tendon constructs and functional spine units are the tissue types used for this study (Figure 1a,c). These samples were chosen due to the existence of different bone structure and soft-tissue attachments and are common regions of interest chosen for bone research. Tissues were harvested from skeletally mature Rambouillet cross ewes (4+ years old) that were euthanized for unrelated studies. All animal studies were performed at Colorado State University, Fort Collins, CO and conducted with approval from the Institutional Animal Care and Use Committee.

Figure 1.

Photographic images of ovine bone samples and corresponding stained tissue sections from the decalcified samples. (a) Gross ovine humeral head (HTC) (triangles) with infraspinatus tendon (arrows). (b) HTC decalcified tissue section. H&E. (c) Gross ovine functional spinal unit (FSU) with end plates (triangles) and disc space (diamonds). (d) FSU decalcified tissue section. H&E. Circles encompass the regions of interest in all images shown in detail in Table 4. Scale bar = 10 mm.

Following gross dissection, all samples were cut at 0.5 cm thick using a diamond blade bone saw (Dorn and Hart MicroEdge, USA). The HTCs were 5 cm long with average width of 2 cm and FSUs were 2.5 cm long with average width of 2 cm. Each sample came from a different animal and there were no serial bone slabs. Tissues were placed in 10% neutral buffered formalin (NBF) at a 10:1 ratio by volume, at RT. NBF was changed to fresh solution every 48 h for 6 days (± 1 day). After fixation, samples were rinsed in running tap H₂0 for 20 min, then placed in a decalcifying solution.

This study included two individual phases as outlined in Table 1. Phase I investigated the effect of six (n = 6) decalcifying solutions on the decalcification time at RT and resulting impact on soft tissue. After Phase I, Phase II subsequently investigated the effect of four (n = 4) decalcifying solutions with two temperatures on decalcification time and resulting impact on soft tissue.

Table 1.

Decalcifying solutions for Phase I and Phase II of the study for ovine humeral head/infraspinatus tendon and functional spinal units.

Ovine Specimen	Phase I Decalcifying Solutions RT With Agitation	*Phase II Decalcifying Solutions 30°C and 37°C With Agitation
Humeral head/Infraspinatus tendon (HTC)	8% TFAA	8% TFAA
	20% TFAA
	8% Formic Acid
	20% Formic Acid	20% Formic Acid
	Formical-4	Formical-4
	XL-CAL
Functional Spinal Unit (FSU)	8% TFAA	8% TFAA
	20% TFAA
	8% Formic Acid	8% Formic Acid
	20% Formic Acid	20% Formic Acid
	Formical-4
	XL-CAL

TFAA Trifluoroacetic acid. RT room temperature

^*Results from Phase I test determined the decalcifying solutions used for Phase II testing.

Phase I solutions included 8% TFAA, 20% TFAA (04901–500, Fisher Scientific, USA), 8% formic acid, 20% formic acid (A119P-4, Fisher Scientific, USA), Formical-4 (1214, StatLab, USA), and XL-Cal (XLC025, Cancer Diagnostics, USA) (Table 1).

For each acid, (n = 6) HTC’s and (n = 7) FSUs were harvested, fixed, and placed in individual, labelled containers with 60 ml of each decalcifying solution at RT and with constant agitation. Decalcification progress was checked daily via radiographic imaging (Ecotron EPX-F2800, Diagnostic Imaging Systems, USA), followed by replacement of the decalcification solutions. Decalcification was considered complete when there were no remaining calcium deposits in bones as evidenced by imaging. Upon decalcification completion, samples were rinsed in running tap H₂0 for 20 min, and placed in SupaMega cassettes (70,065, Electron Microscopy Sciences, USA). Samples were then processed on an automatic tissue processor (Tissue Tek VIP E300, Sakura, USA) using the schedule shown in Table 2, and embedded in paraffin (39,602,004, Surgipath Paraplast Plus, Leica Biosystems, USA). Sections were cut at 5 μm thickness on an automated rotary microtome (Leica RM2255, Leica Biosystems, USA), with high profile, disposable blades (DuraEdge, 7310, Crescent, USA). Sections were placed on Adhesion Superfrost Plus, 50 mm x 75 mm x 1 mm, slides (5057, Brain Research Laboratories, USA). Sections were stained with non-acidified Harris Hematoxylin (6,765,001, ThermoScientific, USA) and Eosin–Phloxine (71,304, ThermoScientific) using the procedure shown in Table 3 and cover slipped using 48 mm x 60 mm no. 1 thickness cover glass (4860–1D, Brain Research Laboratories). Sections were graded by a board-certified veterinary pathologist who was blinded to decalcification methods. Histological interpretation for loss of cellular detail and morphology was scored using a 5-point scale along the following rubric: 1 no discernable effect of decalcification, 2 minimal loss of cellular detail, 3 mild loss of cellular detail, a decrease in section quality but acceptable for histological interpretation, 4 moderate loss of cellular detail with diffuse eosin over-staining and an appreciable decrease in cellular features, and 5 uninterpretable section with marked loss of cellular detail and section quality (Table 4). Osseous and soft tissue regions of the samples were graded separately. Scoring data is not shown and available upon request.

Table 2.

Decalcified ovine bone processing schedule on Sakura VIP E300.

Solution	Number of Changes	Time per change	Temperature (°C)	P/V Agitation
70% ETOH	1	4 h	37°C	On
80% ETOH	2	4 h	37°C	On
95% ETOH	2	4 h	37°C	On
100% ETOH	3	4 h	37°C	On
Xylene	2	4 h	37°C	On
Paraffin	4	4 h	60°C	On

ETOH: ethyl alcohol. P pressure; V vacuum.

Table 3.

Hematoxylin and Eosin staining procedure for decalcified ovine bone.

Step	Time
Bake sections at 60°C	60 min
Deparaffinize in xylene x 2	2 min/change
Rehydrate sections in 100% ETOH x 2, 95% ETOH x 1, 70% ETOH x 1, and DIH2O.	2 min/change
Non-acidified Harris Hematoxylin	10 min
Tap H2O rinses x 3	10 sec, 10 sec, 2 min
1% acid alcohol*	5 sec
Rinse in tap H2O	10 sec
0.25% ammonia H2O	5 sec
Tap H2O Rinse	10 sec
70% ETOH, 95% ETOH	10 sec/change
Eosin Y-Phloxine	15 sec
95% ETOH x 2	1 min/change
100% ETOH x 2	10 sec/change
Xylene x 2	10 sec/ change

^*1% hydrochloric acid in 70% ETOH.

Table 4.

Photomicrographs illustrating loss of cellular detail using semi-quantitative scoring rubric for ovine shoulder (HTC) and spine (FSU) on H&E sections.

Scoring is on a scale of 1 (no loss of cellular detail/best cellular detail) to 5 (severe loss of detail/poorest cellular detail). HTC: intersection of humeral head (shoulder) and infraspinatus tendon. FSU: Spine at the intersection of endplate and disc. Images come from regions of interest (ROI) seen in Figure 1. Scale bar = 0.25 μm.

To create a metric that balanced both decalcification time and resulting cellular detail, a weighted score was created by multiplying the cellular detail score by the days taken to decalcify (Weighted Score = Cellular Detail x Days to decalcify). The two decalcification solutions for HTC’s and FSU’s, which had with adequate cellular detail and the lowest weighted score were carried on to Phase II.

After Phase I results were analyzed using the weighted score, 20% formic acid and Formical 4 were carried forward for the HTC’s and 8% formic acid and 20% formic acid were carried forward for the FSUs to Phase II. 8% TFAA was used in Phase II as a baseline for both HTC’s and FSU’s because this was the in house method of decalcification. Phase II tested these solutions at 30°C and 37°C with agitation. It was hypothesized that increasing the temperature would decrease the time to demineralize the specimens Decalcification procedures otherwise were carried out following the same procedure and scored by the same board-certified veterinary pathologist as described for Phase I. It was determined that only those solutions which yielded samples with an average loss of cellular detail score of ≤3 would be acceptable for research histology purposes.

Results

Phase I results

For the HTC samples, 8% TFAA for 4.8 d, 20% TFAA for 5.8 d, and XL-CAL for 5.2 d were capable of demineralizing the samples relatively fast. Unfortunately, these acids produced mild to moderate loss of cellular detail in the HTC, negatively impacting rf interpretation of the sections. Both formic acid concentrations (8% for 9.6 d and 20% for 7.2 d) and Formical-4 for 11.5 d yielded sections with no to minimal loss of cellular detail in the HTC. Unfortunately, the samples took appreciably longer to decalcify the bone. The two solutions with lowest weighted scores that proceeded to Phase II for HTC’s were 20% formic acid (weighted score 10.3), and Formical-4 (weighted score 16.0). Results for decalcification time and loss of cellular detail are presented in Figure 2.

Figure 2.

Scatterplot illustrating the Phase I results for loss of cellular detail in ovine humeral head (HTC) and ovine functional spinal unit (FSU). Increasing score is shown as 1 to 5 (y-axis) and decalcification time in days (D) on (x-axis). Vertical arrow indicates increasing loss of cellular detail with score 1 with least loss and score 5 with greatest loss, Error bars indicate standard deviation for the decalcifying solutions: 8% and 20% trifluoroacetic acid (TFAA), 8% and 20% formic acid, Formical 4 and XL-Cal.

Similar results were found for Phase I FSU sample testing. For decalcification, 8% TFAA for 3.6 d, 20% TFAA for 3.8 d and XL-CAL for 4 d decalcified the FSUs faster. However, these acids produced sections with unacceptable loss of cellular detail and resolution. The 8% and 20% formic acid and Formical-4 solutions yielded histology with minimal loss of cellular detail. Both 20% TFAA and XL-CAL did not maintain adequate cellular detail (average detail score >3) and did not proceed to Phase II. Out of the remaining solutions, the 8% and 20% formic acid solutions had the lowest weighted score and thus proceeded to Phase II testing for the FSU.

Phase II results

It was hypothesized that increasing the temperature would decrease the time to demineralize the specimens. Interestingly, while for some solutions this decrease in time did occur, a Kruskal–Wallis test (Minitab 18.1, State College, PA) revealed that temperature did not play a statistically significant role in decalcification time (p = 0.450 for the HTC and p = 0.140 for the FSU). However, the two increased temperatures (30°C and 37°C) had a negative impact on cellular detail. A one-way ANOVA with a Tukey post-hoc test revealed that decalcification at RT had statistically significant better cellular detail as compared to decalcification at 30°C and 37°C for HTC samples (p = < 0.010) and for FSU samples, 37°C (p = 0.030) as compared to RT). All HTC samples decalcified at 37°C yielded unacceptable loss of cellular detail. Phase II results are shown in Figure 3

Figure 3.

Scatterplot illustrating the Phase II results for loss of cellular detail in ovine humeral head (HTC) and ovine functional spinal unit (FSU) using three decalcifying solutions and three temperatures for each solution. Increasing score is shown as 1 to 5 (y-axis) and decalcification time in days (D) on (x-axis). Vertical arrow indicates increasing loss of cellular detail with score 1 with least loss and score 5 with greatest loss. Error bars indicate standard deviation for the decalcifying solutions used for HTC 8% trifluoroacetic acid (TFAA), 20% formic acid, and Formical 4 and for FSU, 8% TFAA, 8% and 20% formic acids. Decalcification for each solution was carried out at RT, 30°C and 37°C for all Phase II samples.

Discussion

The purpose of this study was to determine a method of decalcification for ovine bone tissues that optimized decalcification time while minimizing impact on cellular detail and morphology. Specifically, two types of ovine bone that contained tendon and spinal disc tissue attachments were tested to determine both the optimal demineralization solutions and temperature. Upon analysis of the results, it is clear that 20% formic acid at 30°C yielded an optimal decalcification rate and maintenance of cellular detail for the functional spine units. For the humerus-infraspinatus tendon constructs, 20% formic acid at RT proved to be the ideal solution.

It has been shown that 8% TFAA could increase demineralization speed of ovine long bone and FSU samples [14]. However, upon using this acid for HTC samples cellular detail of the tendon is vital, it was evident that 8% TFAA resulted in unacceptable cellular detail. These data reveal the necessity to optimize the decalcification solutions and conditions according to the type of bone to achieve optimal decalcification results and subsequent histologic samples.

Conclusion

In conclusion, optimized decalcification procedures for humerus tendon complex samples and functional spine unit have been outlined. A solution of formic acid at 20% concentration was optimal for both bone types, with decalcification at RT being optimal for HTC samples and at 30°C for the FSU samples.