Enhancing CT imaging: A safe protocol to stain and de‐stain rare fetal museum specimens using diffusible iodine‐based staining (diceCT)

Abstract

Computed tomography (CT) scanning is being increasingly employed in the study of natural history, particularly to investigate the internal anatomy of unique specimens in museum collections. Different techniques to enhance the contrast between tissues have been developed to improve the quality of the scans while preserving the integrity of these rare specimens. Diffusible iodine‐based contrast enhanced computed tomography (diceCT) was found to be particularly effective and reversible for staining tissues in formalin preserved specimens. While it can also be effectively employed to stain ethanol‐preserved specimens of small size, the reversibility of this process and the applicability to large‐bodied animals has never been thoroughly tested. Here, we describe a novel diceCT protocol developed to stain and de‐stain ethanol‐preserved prenatal specimens of baleen whales (Mysticeti, Cetacea). These large (10–90 cm in length only considering early fetal stages) specimens present unique challenges as they are rare in collections and irreplaceable, therefore it is imperative to not damage them with the staining process. Before trying this protocol on baleen whales’ specimens, we conducted a pilot study on commercially available fetal pigs using the same parameters. This allowed us to optimize the staining time to obtain the best results in CT scanning and to test first‐hand the effect of de‐staining on ethanol‐based specimens. External coloration of the specimens is also a concern with iodine‐staining, as stained specimens assume a bright red color that needs to be removed from both internal and external tissues before they can be stored. To test the reversibility of the stain in ethanol‐preserved specimens with fur, we also conducted a small experiment using commercially acquired domestic mice. After these trials were successful, we applied the staining and de‐staining protocol to multiple fetal specimens of mysticetes up to 90 cm in length, both ethanol‐ and formalin‐preserved. Specimens were soaked in a solution of 1% pure iodine in 70% ethanol for 1–28 days, according to their size. After scanning, specimens are soaked in a solution of 3% sodium thiosulfate in 70% ethanol that is able to completely wash out the iodine from the tissues in a shorter time frame, between a few hours and 14 days. The same concentrations were used for formalin‐preserved specimens, but DI water was used as solvent instead of ethanol. The staining technique proved particularly useful to enhance the contrast difference between cartilage, mineralized bone, teeth, and the surrounding soft tissues even when the specimens where scanned in medical‐grade CT scans. The specimens did not suffer any visible damage or shrinkage due to the procedure, and in both the fetal samples and in the mice the color of the stain was completely removed by the de‐staining process. We conclude therefore that this protocol can be safely applied to a variety of ethanol‐preserved museum specimens to enhance the quality of the CT scanning and highlight internal morphological features without recurring to dissection or other irreversible procedures. We also provide tips to best apply this protocol, from how to mix the solutions to how to minimize the staining time.

We describe a novel diceCT protocol developed to stain and de‐stain ethanol‐preserved large museum specimens and test it on foetal pigs and baleen whales, as well as domestic mice. Staining proved useful to enhance the contrast difference different tissues. The specimens did not suffer any visible damage, and in all samples the color of the stain was completely removed by de‐staining. We provide tips to best apply this protocol, from how to mix the solutions to how to minimize the staining time.

1. INTRODUCTION

High‐resolution X‐ray computed tomography (CT) as well as medical CT scanning has been applied increasingly to study both recent and fossil biological specimens. Recent specimens present unique challenges such as tissue preservation, specimen size and density contrast between tissues in CT images (Faulwetter et al., 2013; Giribet, 2010; Rawson et al., 2020). Several methods to enhance resolution of non‐mineralized animal tissues in CT images have been widely used in recent years, mostly by adopting techniques to stain tissues selectively. Some of the most commonly used staining agents are osmium, phosphomolybdic acid (PMA), phosphotungstic acid (PTA), mercuric chloride (HgCl₂), and barium (Gignac & Kley, 2018; Gignac et al., 2016; Metscher, 2009; Pauwels et al., 2013). While all these agents can effectively increase the contrast resolution among tissues, concerns exist due to their potential toxicity, their cost and availability as well as the possibility to effectively remove the stain without damaging the specimens. For example, osmium forms a volatile and toxic compound when exposed to air (osmium tetroxide, OsO₄) (Johnson et al., 2006; Ribi et al., 2008), and PTA, while it can be removed effectively, the process involves sokiang the specimens in a solution with elevated pH that has been shown to damage interal tissues in some cases (Schmidbaur et al., 2015). Additionally, these stains have mostly been tested on small specimens with a maximum volume of a few hundred cm³, and therefore their efficacy and the effects of prolonged staining in larger specimens remain unkown (Gignac et al., 2016; Metscher, 2009; Pauwels et al., 2013). Thus, recent research has focused on finding alternative staining solutions that could provide at least the same level of resolution and that is non‐toxic, reversible, and applicable to a wider range of specimens (Gignac et al., 2016; Metscher, 2009).

Diffusible iodine‐based contrast enhanced computed tomography (diceCT) was found to be particularly effective and easy to perform in multiple studies with fresh and formalin‐fixed specimens (Metscher, 2009). The technique, which uses a solution 1%–2% w/v Lugol's iodine solution (I₃ K) in water, is especially useful to visualize soft tissues such as muscles and neural tissues in mammalian embryos (Gignac & Kley, 2014; Herdina et al., 2015; Wong et al., 2013). Elemental iodine (I₂E) can also be used to stain ethanol‐preserved specimens, as it is soluble in this medium (Gignac & Kley, 2014; Gignac et al., 2016; Metscher, 2009). While performing staining using an ethanol solution is less common in the literature (e.g. Akkari et al., 2015; Early et al., 2020; Hedrick et al., 2018; Herdina et al., 2015) this technique was shown to preserve the difference in grayscale values observed in the CT images between bone, cartilage, and soft tissues, and to actually increase this contrast difference when compared to staining using water‐based Lugol's solution (Li & Clarke, 2015; Li et al., 2016).

Staining specimens with potassium iodide (I₂KI) is known to be reversible, given that sodium thiosulfate (Na₂S₂O₃) can break the molecular bond between iodine molecules and therefore make the characteristic red‐brown color of the iodine stain disappear from the solution and the specimen (Gignac et al., 2016; Schmidbaur et al., 2015). However, the effects of this de‐staining process on the specimens have not been reported in many cases, and some concerns exist that this process might damage the specimens (Gignac & Kley, 2018). De‐staining using sodium thiosulfate has never been applied to samples preserved in ethanol for long periods of time, like museum specimens, as the salt is not soluble in this medium and therefore it would be necessary to immerse them in a water‐based solution that could deform the tissues. These specimens are usually de‐stained by simple diffusion by immersing them in 96% ethanol (EtOH) and frequently changing the fluid before returning them to long‐term storage in 70% EtOH (Early et al., 2020; Gignac et al., 2016; Schmidbaur et al., 2015).

A potential drawback of applying diceCT, especially to unique museum specimens, is the possible shrinkage due to the staining and de‐staining process that might alter the position of and deform soft tissues, making the results of morphological studies less reliable (Hedrick et al., 2018; Vickerton et al., 2013). The staining in a water‐based iodine solution has also been shown to cause bone demineralization, although this phenomenon was less severe using an ethanol‐based solution (Early et al., 2020). Moreover additional problems caused by de‐staining solution are the crystallization of the sodium thiosulfate that may lead to consequent rupture of the tissues of the specimens, as well as bone decalcification caused by this salt (Gignac & Kley, 2018; Gignac et al., 2016). However, these issues are likely negligible in most cases. Multiple studies have demonstrated that the majority of the shrinkage, deformation and demineralization is likely attributable to long‐term preservation in ethanol rather than to the diceCT protocol, as long as the solution concentrations are kept low (≤10% w/v), immersion time in the solution is relatively brief and an ethanol‐based solution is used (Early et al., 2020; Gignac et al., 2016; Hedrick et al., 2018; Vickerton et al., 2013). The relative position of the tissues has also been shown to be preserved when compared to results obtained with traditional dissection (Cox & Faulkes, 2014). The specimens can also be used for histological studies after de‐staining (Hopkins et al., 2015; Nasrullah et al., 2017).

Gignac et al. (2016) provided a thorough review of the available studies in the literature that employed diceCT. Most of the studies have been conducted on small‐bodied animals such as mice or else dissected body parts, with the biggest stained specimens being 5‐10 cm in length. Although, size is likely the most important factor to be taken into account when applying iodine staining (Gignac et al., 2016). This remains the case even in more recent years, with most of the studies applying diceCT having focused on reptiles (e.g. Gignac & Kley, 2018), birds (e.g. Watanabe et al., 2018, Early et al., 2020) and only on a few mammals, mostly rodents (e.g. Nasrullah et al., 2017) and bats (e.g. Hedrick et al., 2018). Moreover most of the specimens used have a well‐documented preservation history or were stained fresh, and there are not many reports of using the technique on large or historical museum specimens. In view of the previous research, we attempted to apply this relatively new technique to the study of embryos and large fetuses of baleen whales (Cetacea, Mysticeti) to provide a template for future studies on museum specimens of similar size.

Very little is known about baleen whale development, especially the prenatal transition between fetal tooth buds and the baleen plates, which are the characteristic keratinized feeding structures that grow from the palate of mysticetes (see Berta et al., 2016, for further discussion). Most of the published information comes from traditional dissection studies (e.g. Ridewood, 1923), with few modern studies employing CT scanning or histology (e.g. Hampe et al., 2015; Thewissen et al., 2017). The tooth buds in fetal specimens are particularly difficult to visualize using CT scanning given the intramembranous ossification of most of the skull bones and the fact that the tooth buds appear to be particularly small in baleen whales. Moreover these specimens have been preserved in various fluid media for 50 to over 100 years and the conditions of the internal tissues are not as ideal for scanning like in fresh specimens (Hampe et al., 2015). The use of diceCT can increase the detail of morphological analysis on these specimens as it enhances the contrast difference between cartilage and bone (Li et al., 2016), which is of utmost importance when analysing developmental sequences, and also helps identifying the presence of developing tooth germs (Nasrullah et al., 2017). This technique can potentially also highlight the anatomy of the developing neural system, heart, and other organs, allowing these scans to be used to investigate other aspects of baleen whale development in the future without needing to dissect these unique specimens.

The baleen whale specimens that we stained represent a series of unique challenges: they are large (from 10 cm to over 90 cm in total length only considering the early fetal stages), uncommon in museum collections, and were mostly collected from the mid‐1800 s to the 1970 s. Fresh specimens are rare and are largely unobtainable given the endangered status of several species of this mammal group (IUCN, 2019) and the strict regulations on their hunting and collections governed by the International Whaling Commission (2018). Most of the museum specimens used had been stored in different water‐based media such as formalin of unknown concentration for many years before being moved to 70% EtOH. Given these changes in preservation medium and the long‐term storage in ethanol, shrinkage and bone decalcification should not be significantly increased by the ethanol‐based diceCT protocol (Early et al., 2020; Fox et al., 1985; Hedrick et al., 2018; Vickerton et al., 2013).

Due to the rarity of fetal mysticete specimens in museum collections, we conducted a pilot study using commercially available fetal pigs applying the same parameters that we would later apply to the baleen whale specimens. In this way, we were also able to observe any enhancement of contrast in the CT images before and after the staining, as well as try the de‐staining process in an ethanol‐based solution firsthand. One concern of museum curators and collections managers is the permanency of the stain, not only within internal tissues, but staining of the external surfaces, including skin and hair. To this end, we also conducted a small experiment to test the reversibility of external staining of the integument using domestic mice.

We then proceeded to apply the staining and de‐staining methods to multiple fetal specimens of mysticetes up to 90 cm in length, preserved in both 70% EtOH and formalin, from numerous institutions in the US and Japan. Even when scanning the specimens using a medical grade CT scanner, which is usually more readily available than a high‐resolution X‐ray CT scanner, we found a consistent improvement in density contrast between soft tissues and cartilage and mineralized bone. Here we present the protocol used to obtain these results and to correctly de‐stain the specimens, in the hope to provide guidelines for future use of this methodology on large mammalian fluid‐preserved museum specimens.

2. MATERIALS AND METHODS

2.1. Specimens

Three pig fetuses and four domestic mice specimens were chosen for the initial testing of the staining and de‐staining procedure. Both species were selected because they are common laboratory animals and are readily available from biological supply companies. Pig fetuses lack body hair and can be purchased at different gestational ages and body lengths, allowing us to simulate the different sizes and skin texture of fetal cetaceans (Table 1, Table S1). Domestic mouse specimens with white fur were selected in order to better visualize the effects of de‐staining on the external appearance of the specimens (Table 2, Table S1).

TABLE 1.

List of specimens used to test staining and de‐staining protocol for fetal ethanol‐preserved specimens, with details on solutions and exposure time. Additional information of these specimens in Table S1. Details on solutions in Table 3

Specimens	Total length (cm)	Staining solution	Staining time	De‐staining solution	De‐staining time
Pig1 (domestic pig)	16.5	1% iodine+95% ethanol	1 week	70% ethanol	5 months (control)
Pig2 (domestic pig)	21	1% iodine+95% ethanol	2 weeks	1% sodium thiosulfate+70% ethanol	3 weeks
Pig3 (domestic pig)	14	1% iodine+95% ethanol	1 week	1% sodium thiosulfate+70% ethanol	1 week
Hump1 (humpback whale)	70	1% iodine+70% ethanol	3 weeks	3% sodium thiosulfate+70% ethanol	2 weeks

TABLE 2.

List of specimens used to test staining and de‐staining protocol for fur, with details on solutions and exposure time. All mice are 9 cm in length. Additional information of these specimens in Table S1. Details on solutions in Table 3

Specimens	Staining solution	Staining time	De‐staining solution	De‐staining time
Mouse1 (house mouse)	None	n/a	None (unstained control)	n/a
Mouse2 (house mouse)	1% iodine+95% ethanol	2 weeks	1% sodium thiosulfate+70% ethanol	2 weeks
Mouse3 (house mouse)	1% iodine+95% ethanol	2 weeks	70% ethanol	2 weeks
Mouse4 (house mouse)	1% iodine+95% ethanol	2 weeks	DI water	2 weeks

The fetal pigs and mice were originally packaged in the supply company's proprietary aqueous solution and we soaked them in 70% ethanol for at least one week before staining in order to simulate the preservation of museum specimens. Each mouse specimen was photographed at each stage of the staining and de‐staining process in order to document the external conditions of each specimen. Because we were interested in the effect of the staining on internal tissues of the pigs, those specimens were CT scanned both prior to and after staining. After the pigs were de‐stained, the specimens were dissected to determine if the stain remained on the internal organs. The pigs were placed in 70% ethanol for long‐term storage at San Diego State University (SDSU) and the mouse specimens were discarded.

We tested the procedure on a fetus of humpback whale (Megaptera novaeangliae) from the San Diego Natural History Museum (SDNHM 25552 – Table 1, Table S1). As with the pig specimens, the humpback whale was CT scanned both prior to and after the stain. Following this trial, the protocol was also applied to other six prenatal specimens of various baleen whale species preserved in 70% ethanol, including two preserved in a water‐based formalin solution (Table S1; Lanzetti, 2019; Lanzetti et al., 2020).

2.2. Staining and de‐staining protocol

Following the protocols of Metscher (2009) and Gignac and Kley (2014), we devised a protocol for creating a consistent iodine stain in an ethanol solution, as well as a sodium thiosulfate ethanol‐based de‐staining solution. While elemental iodine (I₂E), commonly sold in solid crystal form, is soluble in 95% ethanol, sodium thiosulfate (Na₂S₂O₃) crystals are only soluble in water. Our goal was to create a stable solution for both substances and keep the ethanol content around 70% as that is the concentration at which most museum specimens are kept for long‐term storage. Additionally, we stained and de‐stained mysticete specimens fixed in formalin by adapting this procedure to a water‐based solution. Quantities and compounds used for these solutions are listed in Table 3.

TABLE 3.

Staining and de‐staining solutions details. For each type of solution (staining and‐de‐staining) the weight of the solute and volume of the solvent needed to obtain the target solution concentration in 1 L of the respective medium is reported

Staining solutions	1% w/v iodine (I2)
Medium	Metal iodine (g)	Potassium iodide (g)	95% ethanol (L)	DI water (L)
95% ethanol	1	n/a	1	n/a
70% ethanol	6.72	n/a	0.737	0.263
DI water	2.38	10	n/a	1

De‐staining solutions	3% w/v sodium thiosulfate (Na2S2O3) or sodium sulfite (Na2SO3)
Medium	Sodium thiosulfate/sulfite (g) in DI water	95% ethanol (L)	DI water (L)
70% ethanol	30	0.737	0.263
DI water	30	n/a	1

2.2.1. Pigs

To best prepare the iodine stain, it is important to crush the iodine crystals with the aid of a mortar of appropriate material (e.g. glass) before submerging them in ethanol. Metal iodine is a very corrosive and potentially hazardous compound. It must be kept in the dark, under a fume hood and can only come into contact with glass instruments and containers since it will corrode most plastics and metal. The initial staining solution that was used for all pig specimens comprised 1% metal iodine dissolved in 95% ethanol. Given that the pigs were not fixed in alcohol, we decided not to make a 70% ethanol solution for the stain. Pig1 was stained in the solution for 1 week and was then soaked in 70% ethanol in order to observe if the staining solution could be washed out just by diffusion without any additive. Pig1 was then CT scanned after two months along with the other pigs. Pig2 was stained for 2 weeks while Pig3 for 1 week, because of their different sizes and to determine the correct timing for maximizing the effect of the stain on the CT images without damaging the specimens. When the staining was complete, the pigs were transferred into 70% ethanol for a few days and CT scanned along with Pig1.

After all the pigs were CT scanned, we decided to prepare and test a de‐staining solution on Pig2 and Pig3, while we kept Pig1 in 70% ethanol as a control in order to keep observing the diffusion of the stain in ethanol. The de‐staining solution we initially prepared was 1% sodium thiosulfate dissolved in 70% ethanol. This salt has been known to remove the iodine from tissues stained with the Lugol's solution but has never been applied to ethanol‐fixed specimens. To obtain the final concentration, we first prepared a 3.5% solution of sodium thiosulfate in deionized (DI) water, and then we mixed this solution with 95% ethanol to obtain the final 70% ethanol solution at 1% w/v of sodium thiosulfate. Pig2 was left in the de‐staining solution for 3 weeks, and Pig3 for 1 week, checking its skin color frequently in order to observe any changes in the intensity of the stain. Afterward, we soaked the pigs in 70% ethanol and then dissected all three pigs to check their internal coloration along with the external changes.

2.2.2. Mice

To observe how the stain and de‐stain solutions affect specimens with hair, the mouse specimens were individually placed in separate 8‐ounce jars that were filled with different solutions (Table 2). Mouse1 was left unstained to serve as a control, and the remaining mice were placed in the staining solution. Following 2 weeks of staining, each mouse was placed in a variety of solutions for de‐staining (Table 2), except for the unstained control (Mouse1), which was stored in 70% ethanol for the duration of the experiment. The mouse specimens remained in their various de‐staining solutions for two weeks.

2.2.3. Humpback whale

Following the pilot studies with the pigs and mice, we applied the same protocol to stain the humpback whale fetus (Hump1) after it was CT scanned. Since the specimen had been in long‐term storage in 70% ethanol for well over 50 years, we also brought the staining solution to the same concentration by making an initial 1.36% iodine solution in 95% ethanol and then adding DI water. The specimen was left in the staining solution for 3 weeks before being wrapped with ethanol‐soaked cheesecloth and placed in vacuum‐sealed plastic bags to be shipped for a second round of CT scanning in Texas (see below).

After the specimen arrived back at SDSU, the specimen was soaked in the de‐staining solution. Noting the size of the specimen and the relative time it took to de‐stain the pig specimens, we prepared a 3% sodium thiosulfate solution in 70% ethanol. We employed the same procure as we did for the pigs but prepared a 11.4% sodium thiosulfate solution in DI water before adding 95% ethanol. A higher concentration of sodium thiosulfate ensures faster de‐staining, reducing the time of immersion in the solution at minimum and minimizing potential adverse effect on the specimen (Gignac & Kley, 2018; Gignac & O’Brien, 2016; Vickerton et al., 2013). After 2 weeks, we did not observe any further discoloration of the solution and therefore we transferred the specimen to a 70% ethanol solution for return to the museum.

2.3. CT scanning

The pig samples were CT scanned UCSD Medical Center – Thornton Hospital in La Jolla, CA, using a GE Medical System CT scan both before and after the staining (Table 4). In both scans, the pigs were scanned together on the same tray. Pig1 was not scanned before staining as it served as the control. Its size and gestational age, and consequently its ossification level, are comparable to Pig3.

TABLE 4.

CT scan details for fetal ethanol‐preserved specimens

Before staining
Specimen	Voltage (kV)	Current (mA)	Number of slices	Voxel size (mm3)	MorphoSource DOI
Pig2	120	239	104	0.5195 × 0.5195 × 2.5	https://doi.org/10.17602/M2/M166684
Pig3	120	239	104	0.5195 × 0.5195 × 2.5	https://doi.org/10.17602/M2/M166686
Hump1	150	0.11	1880	0.1459 × 0.1459 × 0.1459	https://doi.org/10.17602/M2/M166689

After staining
Specimen	Voltage (kV)	Current (mA)	Number of slices	Voxel size (mm3)	MorphoSource DOI
Pig1	120	253	127	0.7832 × 0.7832 × 3	https://doi.org/10.17602/M2/M166683
Pig2	120	253	127	0.7832 × 0.7832 × 3	https://doi.org/10.17602/M2/M166685
Pig3	120	253	127	0.7832 × 0.7832 × 3	https://doi.org/10.17602/M2/M166687
Hump1	130	0.09	1164	0.2216 × 0.2216 × 0.2216	https://doi.org/10.17602/M2/M166682

The whale fetus (Hump1) was scanned at the University of Texas High‐Resolution X‐ray Computed Tomography Facility in Austin, TX (UTCT) using a NSI CT scanner with a Fein Focus High Power source both before and after the staining (Table 4). There is a significant difference in resolution in the images obtained by these two scanning methods, but we were only interested in observing relative differences in the quality of images provided by the iodine stain for the pigs. We instead scanned Hump1 using a high‐resolution CT scan since we planned to include this specimen in a larger research project on mysticete skull development (Lanzetti et al., 2020). CT scan images for the pigs and the humpback specimens before and after the staining are available on MorphoSource (Table 4).

We did not perform a follow‐up scan after de‐staining due to budgetary restrictions and because it was not the goal of this study to test the efficacy of the de‐staining process on a microscopic level. Similarly, we did not perform histological analyses on the pigs. The main aim of this study was to produce a safe protocol that would allow to apply diceCT to museum specimens, and therefore we were interested in understanding the effects of staining and de‐staining on the external appearance of the specimens, as this can be an issue for displays and long‐term storage. Additionally, it is highly unlikely that any histological or partially destructive studies can be conducted on rare museum specimens such as fetal baleen whales, due to the unique nature of these samples as well as to the fact that their tissues are already frequently altered due to their age and issues in preservation. We encourage future studies to investigate other potential effects of this protocol also on different animals such as birds or reptiles, as well as other curatorial concerns for specimens such as bone decalcification (e.g. Early et al., 2020) where it might be more appropriate.

3. RESULTS AND DISCUSSION

3.1. Staining and de‐staining

3.1.1. Pigs

The skin of the three pig specimens assumed a red color quickly after being placed in the staining solution. Pig2 was kept in the stain for 2 weeks instead of one like the other two specimens given its slightly larger size. This specimen displayed the darkest red color at the time of the scan (Figure 1).

FIGURE 1.

Fetal pigs after staining and de‐staining in lateral view. All pigs were stained using a solution of 1% metal iodine in ethanol. Pig2 and Pig3 were de‐stained using a solution of 1% sodium thiosulfate, Pig1 served as a control and was only de‐stained through diffusion in ethanol. Additional information in Tables 1 and 3

We decided to not use the de‐staining protocol on Pig1, in order to both assess the removal of the stain from internal tissues by ethanol through diffusion alone, and also to have a control to monitor for tissue shrinkage induced by prolonged storage in ethanol (Hedrick et al., 2018; Vickerton et al., 2013). One of the possible downsides of iodine staining and de‐staining is tissue shrinkage but given that the pig samples were fixed in a water‐based solution, we wanted to have a baseline to assess the amount of shrinkage only due to ethanol immersion. Using a 1% sodium thiosulfate solution, 3 weeks were necessary to visually remove the stain color from Pig2 versus the two weeks of staining and the de‐staining solution needed to be changed twice during the process as it had turned orange, showing saturation with iodine. Pig3 instead was completely de‐stained after 2 weeks of immersion in the de‐staining solution without needing to refresh it. This evidence suggests that Pig2 was oversaturated with the stain. The size difference among the pig specimens was not enough to warrant prolonged staining time, and 1 week of treatment was sufficient for the iodine solution to penetrate the skin and diffuse to the internal tissues of the specimens. In fact, when trying to estimate the volume of the specimen by approximating it to a cylinder, where the height is the specimen's total length and the radius is calculated from the maximum circumference, Pig2 (483 cm³) results to be less than twice the size of Pig1 and Pig3 (135–222 cm³), not warranting a doubling of staining time (Table S1). Overstaining can potentially decrease CT image quality and damage the tissues (Gignac et al., 2016; Metscher, 2009). Therefore, it is advisable to keep staining time at a minimum and perform repeated CT scans to check the stain level, if possible, or at least closely visually monitor specimen color during the process.

By visually comparing the de‐stained specimens to the control, we did not detect a visible difference in the shrinkage between these specimens and the control. No samples showed significant differences after the treatment when comparing them to pictures and measurements taken before the staining process and to other fetal pig specimens purchased in the same lot and not subjected to the trials (Figure 1).

To determine the efficacy of the de‐staining process, we conducted a partial dissection of the three pig specimens. In particular, we wanted to observe the coloration and condition of the internal organs and muscles of the abdomen and of the rostral bones in cross section. All internal structures that we could examine in the two de‐stained pig specimens (Pig2 and Pig3) appeared clear from the stain, having re‐acquired a natural coloration compared to fresh samples, and with no visible damage from the procedure. Pig1 also presented a similar coloration to the other samples, as it was taken out of the staining solution about five months before and its ethanol solution regularly changed. Images showing the incision in the abdomen are available in the Supplementary Material (Figure S1).

This trial shows that the iodine stain can be at least partially removed by simple diffusion in ethanol, but this process takes months even for a small specimen and requires continues maintenance and use of high quantities of fresh ethanol solution. De‐staining with sodium thiosulfate is instead quick and effective. While several side effects have been presented for this de‐staining solution such as potential decalcification of bones, tissue tearing and shrinkage (Gignac & Kley, 2018; Gignac et al., 2016), we confirm that this process does not visibly alter tissues, and consistently and quickly removes the stain (Gignac et al., 2016; Nasrullah et al., 2017; Schmidbaur et al., 2015). If both staining and de‐staining solution concentrations and time of exposure is kept low, the specimens will not present permanent damage and they will still be available for further studies and museum displays. Staining time should be estimated based on the size of the specimens, possibly estimating their volume, and de‐staining time should be about half of the time used for staining. Increasing the sodium thiosulfate concentration slightly can reduce the time needed to wash out the stain significantly, potentially mitigating the adverse effects of the de‐staining process (see below).

3.1.2. Mice

Three mouse specimens were soaked in the iodine staining solution for 2 weeks. After staining, the specimens had turned from off white to distinctly yellow (Figure 2). The mice were placed in the de‐staining solution for a total of 2 weeks. After the first week, the sodium thiosulfate solution (containing Mouse2) was yellowed and there was slight discoloration to the ethanol (containing Mouse3). The de‐ionized water (containing Mouse4) experienced no appreciable change in color. All three de‐staining solutions were replaced with fresh solutions after the first week.

FIGURE 2.

Mice before and after staining and de‐staining in dorsal view. All mice were stained using a solution of 1% metal iodine in ethanol. Different de‐staining treatments were used to remove color of the iodine stain from the fur. Mouse1 served as a control and was only de‐stained through diffusion in ethanol. Additional information in Tables 1 and 3

After the full 2‐week de‐staining process, the sodium thiosulfate solution was distinctly yellow, but the ethanol and de‐ionized water were clear and uncolored. The fur of Mouse2 (sodium thiosulfate) retained a faint yellow coloration, but it was not as vibrant or as distinctive as it was in the stained state (Figure 2). The stain appeared to have been removed from the fur of Mouse3 (ethanol) and Mouse4 (DI water) when compared to the unstained control and photographs taken prior to staining.

The discoloration of the sodium thiosulfate solution to a pale yellow during the de‐staining process, coupled with the maintenance of clarity of the de‐ionized water and ethanol, suggests that the iodine staining solution was drawn out of the internal tissues of Mouse2, rather than rinsing clear of the fur. If the stain was removed from the fur only, then one might expect the water and ethanol to discolor. When Mouse2 was briefly rinsed under tap water, the fur noticeably whitened to its original coloration. When Mouse3 and Mouse4 that were treated with ethanol and de‐ionized water were placed in sodium thiosulfate after the initial de‐staining process, the sodium thiosulfate was discolored to a pale yellow in both instances. Furthermore, the fur of Mouse3 and Mouse4 acquired a yellow tint when removed from the sodium thiosulfate. Both mouse specimens were whitened to their initial state after a brief tap water rinse. Thus, the yellow coloration of Mouse2 immediately after de‐staining likely is the result of iodine dissolved in the sodium thiosulfate clinging to the fur, rather than a permanent staining of the fur itself.

3.1.3. Humpback whale

The baleen whale fetus (Hump1) was stained for three weeks following the trials on the pigs and mice. When the specimen was CT scanned, the skin had a consistent red color that had been retained for 3 days, indicating that it was saturated with the iodine solution. This specimen has a considerably large volume of over 6000 cm³ which increased the time needed to reach stain saturation.

To de‐stain Hump1 more quickly, considering its volume and the long staining time, we used a 3% sodium thiosulfate solution. After one week the specimen looked clear of the stain and the solution did not present signs of saturation. Nevertheless, we changed the de‐staining solution and left the specimen in it for one additional week to ensure that no stain was left in the specimen, potentially affecting its long‐term storage.

Although we did not perform dissection or CT scans after the stain was removed, the external appearance of the specimen was overall similar as before the staining process (Figure 3). This specimen, as many cetacean fetal specimens in museum fluid collections, present a yellow coloration likely due to the fixatives used initially on them and long term storage in ethanol (see Lanzetti, 2019 and Lanzetti et al., 2020 for additional examples). We noticed the precipitation of 2–3 mm long sodium thiosulfate crystals in small patches on the skin. The formation of these crystals is likely due to the fact that the salt is not soluble in ethanol. In order to avoid more crystals forming that could possibly cause tears in the tissue, we rinsed the specimen for 15 s under running tap water before putting in into 70% ethanol for long‐term storage. Briefly rinsing museum specimens with water does not affect the tissues since they have been fixed in ethanol for extended periods of time (Simmons, 2014). For example, the humpback whale specimen was collected in 1961, and after fixation in formalin was probably preserved in ethanol for over 50 years. This procedure also prevents worse damage to the tissues since the sodium thiosulfate quickly dissolves in water and can be washed off completely from the skin of the specimen. The presence of the salt crystals as well as the staining process have been shown to not affect the nature of the tissues in a significant way, and specimens after the procedure are still suitable for histological studies (Gignac & Kley, 2018; Hopkins et al., 2015).

FIGURE 3.

Humpback whale fetus before and after staining and de‐staining in lateral view. It was stained using a solution of 1% metal iodine in ethanol and de‐stained using a solution of 3% sodium thiosulfate. The grey area in the Initial State figure covers the specimen's label. The label was then removed to avoid damaging it in the staining process. Additional information in Tables 1 and 3

3.2. CT images and 3D rendering quality comparison

3.2.1. Pigs

Before the stain, only the major mineralized structures were visible clearly in the CT scans of all pig samples, including the ribs, teeth, portions of the skull and of the limb bones, and vertebral column (Figure 4, Video S1). Very little detail of the skull morphology was distinguishable since most ossification centers in this specimen had yet to form. After staining, all major skeletal structures could be distinguished in the pigs, and the head appeared particularly defined in the CT images, with the ocular and brain region clearly highlighted by the stain (Figure 4, Video S1). All of the other internal organs were also visible and even their internal features were distinguishable. This result is particularly important since it shows that our technique highlights details of the anatomy at a comparable level with traditional dissection, making scans obtained with iodine staining suitable for multiple kinds of studies without having to destroy rare specimens. Pig2 appears particularly saturated in the extremities, which is indicative of overstaining as we hypothesized by observing its external color. This indicates a correlation between the external color and the level of diffusion of the iodine and resulting CT imaging quality. Even if not exactly quantifiable, this observation can be used as an overall indication of the time needed for staining the specimen.

FIGURE 4.

Pig2 CT slices without and with iodine staining. (a) CT section of the entire body in lateral view displaying the tooth row without iodine staining, (b) comparable CT section with iodine staining, (c) CT section of the entire body in lateral view through the midline of the body without iodine staining, (d) comparable CT section with iodine staining. In the CT images with the iodine stain (b, d) partially ossified or cartilaginous elements in the rostrum and neurocranium are recognizable in a slightly lighter grey compared to the other soft tissues. In a and c, these elements are not recognizable. The stain also makes it possible to highlight internal organs such as the brain and the intestines. Both sets of images were obtained using the same machine (GE Medical System CT at UCSD Medical Center – Thornton Hospital in La Jolla, CA), details on the scanning parameters in Table 4. Complete image sequence for CT scans before and after for Pig2 available in the Supplementary Material (Video S1). Some important skeletal elements and tooth germs are labeled in the figures. Abbreviations: Mc, Meckel's cartilage; nc, nasal cartilage; sc, scapula; tc, thyroid cartilage; tg, tooth germs

3.2.2. Humpback whale

The differences in contrast and quality of the CT images are subtler for the humpback whale specimen, given that it was scanned using a high‐resolution X‐ray machine, but they are still important to highlight key aspects of the internal morphology. Before the staining, the head region was poorly defined in the CT images, and the rostral bones were mostly indistinguishable from the surrounding soft tissues (Figure 5, Video S2). This may be an effect of decalcification due to fixation in formalin, which is a common problem with old museum specimens (Fox et al., 1985; Hedrick et al., 2018; Simmons, 1995). While it was possible to recognize developing teeth in the jaws, only the larger tooth germs were identifiable in each row. After the staining (Figure 5, Video S2), the contrast between the rostral bones and the surrounding soft tissues is better resolved. It is possible to recognize tooth germs on all four quadrants of the upper and lower jaws, and it is easier to isolate individual germs from the surrounding bone and cartilage. The difference in contrast between the different tissues highlighted by the staining also makes it possible to recognize and reconstruct in 3D all the ossified skull bones. Postcranially, the limb bones, ribs and vertebral centra appear well defined and are visible on both sides of the body. Internal organs such as the brain and the intestines are clearly recognizable as well, which is again an advantage of this staining method and allows these CT data to be used to study the development of other systems in baleen whales besides the mineralized skeleton. This aspect is particularly important when digitizing museum specimens and it is a strong argument in favor of using this technique. By applying diceCT it is possible to obtain high quality images of the internal anatomy of the specimens, which can then be used for a variety of different studies, both looking at hard and soft internal tissues, without needing to use different investigation methods (e.g. magnetic resonance – MRI), stains (e.g. PTA) or destructive procedures in the future.

FIGURE 5.

Humpback whale fetus (Hump1) CT slices without and with iodine staining. (a) CT section of the skull in dorsal view without iodine staining, (b) comparable CT section of the skull with iodine staining, (c) CT section of the entire body in lateral view without iodine staining, (d) comparable CT section of the entire body in lateral view with iodine staining, (e) 3D rendering on the skull in later view before iodine staining, (f) 3D rendering on the skull in later view after iodine staining. As for the pigs, in the CT images with the iodine stain (b, d) partially ossified or cartilaginous elements in the rostrum and neurocranium are recognizable in a slightly lighter grey compared to the other soft tissues. In (a, c), these elements are not recognizable. The quality of the 3D rendering of the skull (e and f) is also affected by these differences. In (f), additional bones (in yellow) and process are recognizable, and their shape is better defined. Nasals (in green) are also visible and can be segmented separately, tooth germs (in red) are visible on all sides of both lower and upper jaws. Both sets of images were obtained using the same machine (NSI CT scanner with a Fein Focus High Power source at UTCT – Austin, TX), details on the scanning parameters in Table 4. Complete image sequence for CT scans before and after for Pig2 available in the Supplementary Material (Video S2). Ossified elements and tooth germs are labeled in the figures. Abbreviations: d, dentary (=mandibular rami); f, frontal; ip, infraorbital process (maxilla); inp, interparietal; m, maxilla; n, nasal; p, parietal; pm, premaxilla; ra‐ul, radius and/or ulna; so, supraoccipital; sq, squamosal; tg, tooth germs

3.2.3. Other protocol applications

Subsequent to the pilot study described here, we stained, CT scanned and successfully de‐stained four additional ethanol‐preserved and two formalin‐preserved pre‐natal mysticete specimens. These specimens greatly varied in volume, from about 200 cm³ to over 20,000 cm³ (Table S1 – see also Lanzetti, 2019, Lanzetti et al., 2020). The largest specimen is at the upper size limit for applying this protocol. After three weeks of staining, a preliminary scan showed that the stain was not penetrated fully and therefore one extra week of staining was necessary. Therefore we do not recommend trying to stain specimens larger than 10,000–15,000 cm³ with the current protocol to avoid exposing them to the staining and de‐staining solution for prolonged amount of time, increasing the risk of tissue damage.

The ethanol‐preserved specimens were subject to the same protocol as the humpback whale trial, while the formalin‐preserved specimens were stained using a water‐based solution with a 1% iodine (I₂) concentration, obtained with a mixture of Lugol's and metal iodine (Table 3). Sodium thiosulfate was not available for use in de‐staining at the Japanese facility, so the de‐staining solution was made using sodium sulfite (Na₂SO₃), a common reagent used in photography. Sodium sulfite has similar properties to sodium thiosulfate and dissolves well in water. We achieved the same visual results for the de‐staining process in a comparable time by using a 3% sodium sulfite solution in de‐ionized water (Figure 6). Although this compound should be tested separately for de‐staining ethanol‐preserved specimens, we report that it is possible to use this more readily available compound for safely de‐staining specimens preserved in water‐based solutions.

FIGURE 6.

Minke whale fetuses before and after staining and de‐staining in lateral view. They were stained using a solution of 1% w/v iodine in DI water and de‐stained using a solution of 3% sodium sulfite. Additional information in Tables 1 and 3

4. CONCLUSIONS

Overall, we successfully developed a safe protocol to stain museum legacy specimens with metal iodine in order to highlight details of the soft tissues that cannot be seen without a contrasting solution and subsequently de‐stain the same ethanol preserved specimens. The enhancement of density contrasts during the CT scanning is important especially in fetal specimens, since the skeletal structure is not completely developed yet and it is very difficult to distinguish the developing bones from the surrounding soft tissues, and in some circumstances, from the cheesecloth and other non‐biological materials surrounding the specimen during scanning.

To help other researchers to apply this technique to different specimens, mammalian or non‐mammalian, and at different ontogenetic stages, in Table 5 we provide a workflow with suggested staining and de‐staining times based on our experience on mammalian samples, as well as some general tips. The staining and de‐staining time should be adjusted depending on the size of the specimens and the permeability of the skin, and to limit exposure we recommend regularly checking the color of the specimens during this process or conducting multiple CT scans to check the internal state of the samples. It is always advised to aim for the lower amount of time possible and check the penetration of the stain with a quick trial CT scan to avoid damaging the specimen with overstaining. Our protocol is safely and effectively applicable to specimens up to 90 cm in length or 20,000 cm³ in estimated volume, although it is not recommended to apply diceCT to larger specimens as it would imply exposing the samples to the reagents for very extended periods of time, possibly more than one month for the staining solution, causing shrinkage and other damage to the tissues. We highly recommend to first test the protocol on either commercially available or replaceable specimens similar to the museum specimens that are the aim of the proposed study. This will help demonstrate to curators that this protocol is safely applicable to the specimens in their care and will also help the researchers determine the best staining and de‐staining time for their particular set of specimens. A follow‐up CT scan after de‐staining could also help quantify the extent of the removal of the iodine stain form the internal tissues on a microscopic scale.

TABLE 5.

Step‐by‐step guide to diceCT protocol for ethanol‐preserved large specimens. The staining and de‐staining time are only given as guidance and tests should be conducted on replaceable or commercially available specimens before applying the protocol to rare museum specimens. The protocol for formalin‐preserved specimens is similar, only ethanol is not used as solvent. Details on solutions in Table 3

Step 1. Prepare staining solution

Mix metal iodine and ethanol to obtain 1% w.v. staining solution. Dissolve metal iodine in 95% ethanol and then add DI water to bring solution to desired concentration. Prepare at least twice the volume of the solution that the specimen is preserved in. Always use glass or hard plastic containers to mix iodine as it is highly corrosive when in pure metal form.

Step 2. Staining of specimen

Immerse the specimen in the staining solution until the external color remains stable. Measure the specimen length and maximum circumference first to estimate the volume as a cylinder. Approximate time needed for staining: 100–1000 cm3 – 12 hr to 7 days 1000–5000 cm3 – 7–14 days >5000 cm3 – 14–28 days For specimens >5000 cm3, the staining solution might need to be completely refreshed at least once after 14 days. Presence of fur or other integument requires additional time to allow the stain to diffuse. Frequently check the specimen to avoid overstaining and aim for the shortest staining time possible. Move specimen to 70% ethanol after staining is complete if CT scanning cannot be performed right away but aim to CT scan within a few days to avoid washing out the stain.

Step 3. Prepare de‐staining solution

Mix sodium thiosulfate and ethanol to obtain 3% w.v. de‐staining solution. Dissolve sodium thiosulfate in DI water and then add 95% ethanol to bring solution to desired concentration. Prepare at least the same volume as the staining solution, plus some extra to clean up spills and benches.

Step 4. De‐staining of specimen

Immerse the specimen in the de‐staining solution until the external color of specimen and solution remains stable. The approximate time to de‐stain the specimen effectively is about half of the time used for staining: 100–1000 cm3–6 hr to 4 days 1000–5000 cm3 – 4 to 7 days  > 5000 cm3 – 7–14 days For specimens >5000 cm3, the de‐staining solution might need to be completely refreshed after 7 days. Presence of fur or other integument requires additional time to remove the stain. Frequently check the specimen to minimize the time spent in the solution and aim for the shortest de‐staining time possible. Before moving the specimen to 70% ethanol for long term storage, check for the presence of salt crystals on the skin and quickly rinse under running water to remove them. This will avoid build up during storage.

While the absorption rate of each specimen can differ based on its preservation status and the characteristics of the organism (e.g. presence of fur, ontogenetic stage, etc.), the iodine is consistently absorbed preferentially by neural tissue, cartilage and teeth, and therefore can highlight these structures relative to the other tissues and organs. This produces a clear improvement of the resolution of the CT images obtained, even when using a medical‐grade CT scanner, the most widely available type of scanner. By applying diceCT, therefore, it is possible to obtain CT images that can then be used to study multiple aspects of the anatomy of the specimens without the need to subject them to additional procedures.

In the future, we hope this technique is applied to more museum specimens of different types, so that this protocol can be adapted to other organisms (e.g. reptiles, birds) and can be used to study different systems (nerves, blood vessels, bones, etc.). Histological studies have already been performed in specimens after iodine staining (e.g. Nasrullah et al., 2017), and it can be used to complement the results obtained using diceCT. DNA analysis of preserved specimens is very difficult to perform, but there are emerging new techniques that might allow researchers to collect genomic data from museum samples (e.g. Derkarabetian et al., 2019), and therefore it would be interesting to try extract DNA from specimens after the diceCT protocol has been applied to them. Unfortunately, fetal baleen whale specimens are extremely rare and likely not preserved well enough for any of these follow up studies to be conducted. Though, after de‐staining, none of the specimens presented visible alterations, and even fur color returned to normal, showing that this technique is suitable at least for specimens that need to be displayed or stored long‐term. As natural history collections become a resource of ever‐growing importance in the future (Hedrick et al., 2020), diceCT can be used a powerful tool as a non‐invasive and reversible method to digitize collections and preserve these specimens with an unprecedented level of detail for future studies.

CONFLICT OF INTEREST

The authors have no conflict of interest to declare.

AUTHORS’ CONTRIBUTIONS

A.L. and E.G.E. designed the study. A.L. performed the experiments on the pigs and baleen whales. E.G.E. performed the experiments on mice. A.L. and E.G.E. elaborated the results and wrote the manuscript.

Supporting information

Fig S1

Table S1

Video S1

Video S2

DATA AVAILABILITY STATEMENT

The data (CT images) that support the findings of this study are available on MorphoSource (DOI listed in Table 4).