Abstract
The Göttingen minipig is being used increasingly in biomedical research. The anatomical structure of the porcine peripheral nervous system has been extensively characterized, but no equivalent to the dermatome map, which is so valuable in human neurophysiological research, has been created. We characterized the medullar segmental skin and muscle innervations of the minipig hind body, using neurophysiological methodology. Six adult minipigs underwent unilateral laminectomy from L2 to S3, exposing the nerve roots. The skin of the hind part of the body was divided into 36 predefined fields, based on anatomical landmarks for consistent reproducibility. We recorded the evoked potential in each exposed nerve root L2‐S3 for cutaneous stimulation of each skin field, mapping the sensory innervation of the entire hind body. We subsequently recorded the motor response in seven predefined muscles during sequential stimulation of the L2‐S3 nerve roots. We obtained a clear sensory evoked potential in the nerve roots during stimulation of the skin fields, allowing us to map the sensory innervation of the minipig hind body. Neurophysiological data from skin stimulation and muscle recordings enabled us to map the sensory innervation of the Göttingen minipig hind body and provide information about muscular innervation. The skin fields were sensory innervated by more than one root. The muscles each had one dominant root with minor contribution from neighboring roots. This is consistent with experimental data from human studies.
Keywords: dermatome, minipig, neurophysiology, pig, segmental innervation, Sus scrofa, swine
Introduction
The Göttingen minipig is increasingly being used in biomedical research as an attractive alternative to other animal models (Gutierrez et al. 2015). The relatively large size of the animal means that standard surgical techniques and surgical equipment designed for humans can be used (Sorensen et al. 2011). For a large experimental animal, minipigs are relatively inexpensive, easy to handle, and economic to house.
In recent years, the pig has found a use in research concerning electrical stimulation of the spinal cord and/or nerve roots (Odenstedt et al. 2011; Hachmann et al. 2013; Solis et al. 2013; Foditsch & Zimmermann, 2016; Guiho et al. 2016; Kowalski et al. 2016; Howard‐Quijano et al. 2017).
In human clinical research, a profound understanding of the spinal and peripheral nervous system is a prerequisite for designing experiments aimed at affecting those structures, and the classic dermatome map of man serves not only as central clinical tool but also as an important reference in scientific experimentation.
The anatomical structure of the porcine peripheral nervous system has been extensively studied and characterized. Detailed descriptions based on dissection studies have been published in the classic veterinary literature (Bosa & Getty, 1969; Sisson et al. 1975; Sack, 1982; Nickel et al. 1986, 1992) but, to our knowledge, no porcine equivalent to the human dermatome map has been created.
In this study, we aimed to characterize the medullar segmental skin and muscle innervations of the hind limb region in the minipig using neurophysiological methodology.
Methods
Animals
Six young adult (age 18+ months) female Göttingen minipigs weighing 30–40 kg (Ellegaard, Dalmose, Denmark) were used in the study.
The minipigs were fed a restricted diet (Special Diets Services, Aarhus, Denmark) with tap‐water available ad libitum. The animals were housed in standard pig pens in an animal room provided with filtered air, a temperature of 21 °C ± 3 °C, and a relative humidity of 55 ± 15%. The room is designed to provide 10 air changes per hour and is illuminated to give a cycle of 12‐h light and 12‐h darkness.
All procedures were approved (license 2008‐561‐1557) by the Danish Experimental Animal Inspectorate, The Danish Veterinary and Food Administration, and were conducted according to Directive 2010/63/EU.
Anesthesia
Before leaving the animal facility, the pigs were sedated with 25 mg midazolam (Hameln Pharma, Hameln, Germany) administered intramuscularly.
Upon arrival at the test facility, anesthesia was induced with an intramuscular administration of a 7 : 3 mixture of S‐ketamine 25 mg mL−1 (Pfizer, New York, NY, USA) and midazolam 5 mg mL−1 titrated to effect. Typical dose was 0.3–0.45 mL kg−1.
Two 22G peripheral intravenous (i.v.) lines were established in the ear veins, and anesthesia was supplemented with i.v. administration of 0.1–0.15 mL kg−1 of the S‐ketamine/midazolam mixture.
Pigs were intubated (Ettrup et al. 2011) using a size 5.5 cuffed endotracheal tube and connected to an MCM 801 ventilator (Dameca, Rødovre, Denmark) supplying a 1 : 1 mixture of O2 and air.
Ventilation parameters were set to a tidal volume of approximately 7 mL kg−1, and the respiratory rate was regulated to ensure a stable end‐respiratory CO2 between 4.0 and 5.5 kPa (note: no data are available on normal ETCO2 for minipigs; Ellegaard, pers. commun.). Oxygenation was continuously monitored and kept at 98–100%.
Anesthesia was maintained with isoflurane (Baxter, Deerfield, IL, USA) 1.5–2.5% and buprenorphine (Beckitt Benckiser, Slough, UK) 0.6 mg intramuscular (i.m.) supplemented with 0.3 mg after 6 h.
To fully expose the nerve roots (see Surgical procedure), it proved impossible to avoid lesioning an extensive venous plexus in the lateral spinal canal, which caused some bleeding. To help control hemorrhage during surgery, 1500 mg of tranexamic acid (Pfizer) was slowly administered intravenously prior to incision. A 700‐mL aliquot of isotonic saline solution was administered during surgery, and blood loss during surgery was replaced thereafter 3 : 1 with isotonic saline.
A size Fr 8 urethral catheter was inserted to monitor urine output and relieve bladder tension during surgery.
After completed surgery and neurophysiological experiments, animals were euthanized with an overdose of 20 mL intravenous pentobarbital 200 mg mL−1 (Veterinærapoteket København Universitet, Copenhagen, Denmark), corresponding to minimum 100 mg kg−1.
Surgical procedure
Animals were shaved and placed in a prone position on a heating blanket with the upper body covered by a blanket to prevent heat loss.
Surgery was performed by an experienced neurosurgeon (J.C.S.). A midline incision was made from the upper part of the sacrum to Th14 and the lamina and facet joints were exposed using a left‐sided conventional subligamentary surgical approach, following the spinous processes. A total left lumbosacral laminectomy was then performed, exposing the spinal dura and left nerve roots from root S3 to L2 (note: the Göttingen minipig, like most common landrace breeds, has 14 thoracic, six lumbar, and four sacral spinal nerves). The position of each lumbar spinous process was marked with ink on the skin for identification of each nerve root for the remainder of the experiment.
After recording of the skin potentials, the fascia covering selected muscles (see Neurophysiology) was exposed by careful removal of skin and subcutaneous fat.
Immediately following sacrifice of the animal, the spinal column was exposed above Th14 (identified by the lowest set of ribs) to verify the labeling of lumbar and sacral nerve roots.
Neurophysiological methods
Sensory evoked potentials
With the aid from an experienced veterinarian (A.K.A.O.) and using a permanent marker, the hind body of the pig was divided into 36 predefined fields. To avoid any inconsistencies caused by the variations in animal size, we standardized the markings of the fields based on easily recognizable anatomical landmarks. A precise description of the markings can be found as Appendix 1.
Two sensory needles [Disposable Sensory Needle Electrode, length 50 mm, diameter 0.70 mm (22G), active recording area 5.0 mm2, Alpine Biomed, Skovlunde, Denmark] were used as recording electrodes. One electrode was carefully inserted directly under the epineurium and the reference electrode was inserted in the tissue in close vicinity of the roots (Fig. 1). The needles were connected to a 4‐channel amplifier of a Dantec Keypoint EMG system v3.25 (Dantec, Skovlunde, Denmark).
Figure 1.
Experimental setup. (A) An example of a recorded sensory evoked potential at the levels of the roots L2‐S3 following stimulation of the skin fields. A full description of the parameters can be found in the Results section sensory evoked potentials. (B,C) Placement of the recording needles in the nerve roots for obtaining sensory potentials. (D) Placement of the recording needles in the exposed muscles for obtaining motor responses.
Two identical needle electrodes were used as stimulation electrodes and inserted subcutaneously in the center of each of the 36 fields in succession. Stimulation was applied to the skin at this location with an increasing stimulus intensity until visible muscle contractions were elicited, and then reduced just below the motor threshold. Thus, the entire field was not stimulated, merely the geometrical center. The approach with varying intensity between fields was chosen to avoid direct muscle‐nerve evoked potential, which would cause noise from muscle activity, as well as to compensate for variations in thickness of subcutaneous fat layer.
We applied 3–5 Hz repetitive electric stimulation supplied from a bipolar stimulator with a square pulse width of 0.2 ms to each of the 36 fields, and sensory evoked potentials were recorded simultaneously from each nerve root L6‐S3 using the averaging technique, with sequence of 200–400 repetitions. The amplifier gain was set at 5 μV per division and was adjusted during the recordings depending on the amplitude of the potential, typical gain being between 2 and 10 μV. All responses were recorded using a sweep duration of 10 ms and a frequency bandwidth of 20 Hz (high‐pass filter) to 3 kHz (low‐pass filter).
After a full recording of all 36 fields, the process was repeated with recordings made from nerve roots L2–L5.
We tested the authenticity of the recorded potentials by measuring the approximate distance between the stimulation and recording electrodes (15–20 cm) and the latency (mean 3.9 ms, range 1.1–6.3 ms), giving us a nerve conduction velocity of approximately 50 m s−1, which is comparable with previous reports on young minipigs (Hort‐Legrand et al. 2006). Moreover, using the averaging technique, we ensured the reproducibility of the recorded potentials.
Motor examination
The nerve roots L2‐S3 were stimulated directly using 1‐Hz repetitive stimulation, employing the sensory needles mentioned above, only now connected to a bipolar electrical stimulator as described above. All responses were recorded using a sweep duration of 2 ms and a frequency bandwidth of 20 Hz (high‐pass filter) to 5 kHz (low‐pass filter). The amplifier gain was set at 2 mV per division and was adjusted during the recordings depending on the amplitude of the potential, the typical gain being between 2 and 10 mV.
Recordings were made from identical needles inserted into six individual, predefined muscles in the hind limb with direct equivalents in the human musculature (Fig. 1B–D). The six muscles were as follows: m. biceps femoris, m. peroneus tertius, m. semitendinosus, m. tensor fascia latae, m. gluteus medius, and m. soleus. Additionally, for m. coccygeus, we observed a characteristic tail lift.
Statistical analysis
For all recordings of sensory potentials, the following parameters were recorded for the response of each of the eight roots to stimulation in each of the 36 cutaneous fields in each of the six pigs: stimulation intensity, latency, amplitude, and repetition count. All data were transferred to an excel spreadsheet (Microsoft, Redmond, WA, USA) for data processing and creation of 3D bar charts.
The central parameter, the amplitude of the evoked potential for each of the 36 fields derived from each of the eight roots L2‐S3, was calculated as a mean for the six pigs. The amplitude was measured from the most positive peak to the most negative peak.
The overall size of the amplitudes, AmplitudeRAW, varied in size between the pigs. To compensate for individual differences in sensitivity between the pigs, the mean amplitude, AmplitudeMEAN, of all responses was calculated for each pig. The amplitude of each potential was then calculated as a relative value (AmplitudeREL = AmplitudeRAW/AmplitudeMEAN) for each pig, and the mean of the resulting AmplitudeREL was then calculated across the six pigs.
For graphical representation of the results (see section below), the following rule was applied: for each skin field, the maximum of the mean AmplitudeREL of all roots was found. The relative contribution of each root on the skin field was calculated as the percentage of AmplitudeREL to the found maximum value. A root was considered level 1 if this number was in the upper quartile (75–100%), level 2 if it was in the 50–75% quartile, and level 3 if it was in the 25–50% quartile.
The intensity of the sensory stimulation could be responsible for the size of the amplitude of the resulting evoked potential. To control for this, a Spearman rank correlation test was performed for intensity vs. amplitude for all potentials where AmplitudeRAW > 0 using stata IC 11 (StataCorp, College Station, TX, USA).
Unlike the sensory evoked potentials where the amplitude could be accurately quantified, the muscle activation was graded only as activation value 0 (no response) or 1 (clearly detectable muscle contraction). The mean was calculated across all pigs, meaning that the resulting value ranged from 0 (none of the six pigs had a clear muscle response from the muscle in question when the root was stimulated) to 1 (all six pigs had a clear muscle response from the muscle in question when the root was stimulated).
Graphical representation
To illustrate the sensory innervation of the skin fields, an anesthetized minipig received a full‐body CT scan. Anesthesia was performed as described for the surgical procedures.
The resulting DICOM files were reformatted with a dedicated 3D volume rendering plug‐in for horos v2.0.1 (OpenSource freeware, https://www.horosproject.org/) with the following settings on a 16‐bit CLUT‐editor: WL/WW: Other, Opacity: Linear Table, Contrast: High and Fine detail. The Cut tool was used to remove nonanatomical graphic elements.
Subsequently, four suitable angles were chosen to present the minipig hind body. Bone structures were presented by changing value and alpha colors in horos 3D reconstruction to identify the anatomical landmarks used in the creation of the skin fields (see Appendix 1).
Images were transferred to a vector graphics program, Adobe illustrator CC 2018 (Adobe Systems, San Jose, CA, USA). Skin fields were marked on the rendered images using a vector pen tool and a Bamboo graphic pen tablet (Wacom, Kazo, Japan). The innervating nerve roots were marked as L2‐S3 according to the results displayed in Fig. 2. Letter sizes represent the relative influence of the individual roots: large font for level 1 roots, medium font for level 2, and small font for level 3 (for description of the levels and calculation methods, see Statistical analysis section above).
Figure 2.
Sensory innervation, data. Bar height denotes the amplitude of the evoked potential for each of the eight tested roots (L2‐S3) for each of the 36 skin fields, shown as the mean for the six pigs. The displayed amplitudes, Mean AmplitudeREL, are calculated as the mean of all six pigs of the evoked potential amplitude relative to the mean evoked potential of each individual pig, AmplitudeREL. See the Statistical analysis section for details.
Results
Sensory evoked potentials
Stimulation of each of the 36 skin fields elicited a significant evoked potential, except for fields 33 and 36, which were likely innervated by the L1 root. Figure 1A shows an example of recorded sensory action potentials from the L2‐S3 root during stimulation. On the x‐axis, time is shown at 2 ms per div (division), and on the y‐axis the amplitude in μV. The first four channels have a gain of 5 μV per div, and for the last four channels, the gain was increased by 2 μV per div. The total number of stimulation for the first four channels was 422, and for the next four channels 642. The latency of the potential is showed on the right, ranging from 2.6 to 4.2 ms, and the peak‐to‐peak amplitude from 0.69 to 12.6 μV. For this recording, the signals obtained from channels 2–5 were stored for analysis.
All fields were sensory innervated by more than one root, manifested as an evoked potential from more than one root. Most skin fields had one or two dominant roots, manifested as a perceptibly higher amplitude on average from the root(s).
The data for the sensory innervation of the skin fields are graphically represented in Fig. 2. The height of the bars represents the mean value for all six pigs of the sensory evoked potential amplitude relative to the mean of each pig's amplitudes, AmplitudeREL.
Figure 3A–D displays the hind body of the Göttingen minipig viewed from four different angles, illustrating the location of each skin field. Marked on the drawings are the innervating roots, based on the same data as Fig. 2. The size of the fonts reflects the relative contribution of each nerve root.
Figure 3.
Sensory innervation, body map. (A) Rear view of the minipig hind body. The table to the left shows each of the 36 skin fields with the nerve roots involved in sensory innervation. Letter size marks the relative contribution of each root (see Statistical analysis section for details). The contributing roots are also marked directly on each visible skin field. Stimulation was performed in the geometrical center of each field, marked with a red dot. (B) Lateral view of the minipig hind body. The contributing roots are marked directly on each visible skin field. (C) Anterior view of the minipig hind body. The contributing roots are marked directly on each visible skin field. (D) Medial view of the minipig hind body. The contributing roots are marked directly on each visible skin field.
The Spearman test showed no correlation between the stimulation intensity and the resulting evoked potential amplitudes (Spearman's rho = 0.1796).
Muscle innervation
We obtained a definite response from the seven muscles investigated with stimulation of the nerve roots, muscles 1–6 with a clear EMG response, and muscle 7 with a characteristic tail lift.
All the seven muscles had one nerve root with an activation value of 1, meaning that for all pigs, stimulation of the same nerve root produced a clear muscle response. The adjacent nerve roots contributed with only minor variation.
These dominant roots were as follows: m. biceps femoris (L6), m. peroneus tertius (L6), m. semitendinosus (L6), m. tensor fascia latae (L5), m. gluteus medius (L6), m. soleus (L6), and m. coccygeus (S3).
The data are summarized and graphically represented in Fig. 4.
Figure 4.
Muscle innervation. Bar height denotes the fraction of the pigs where stimulation of the nerve root elicited a clear muscle response, Activation value. The muscles are 1: m. biceps femoris, 2: m. peroneus tertius, 3: m. semitendinosus, 4: m. tensor fascia latae, 5: m. gluteus medius, 6: m. soleus, and 7: m. coccygeus.
Discussion
In this paper, we present a thorough sensorimotor mapping of the hind body of the Göttingen minipig. To our knowledge, this is the first attempt at mapping the sensory innervation of the pig using electrophysiology.
Our experimental setup enabled us to obtain sound neurophysiological data from skin stimulation and muscle recordings alike, resulting in a mapping of the sensory innervation of the Göttingen minipig hind body.
The surgical approach proved apt for exposing the nerve roots and allowed us to position each recording needle reliably into the nerve root, causing minimal damage to the nerve fibers. The main problem with the surgical procedure were the venous plexuses that were increasingly prone to lesions with more lateral progression of the laminectomy. Besides standard surgical management of the bleeding, pretreatment with i.v. tranexamic acid proved useful in our experience, with the operator noting markedly less hemorrhage. Our dose of 50 mg kg−1 was about five times higher than the recommended dose for human subjects and also high compared with the few reports of its use in pigs (Sondeen et al. 2016; Zentai et al. 2016). However, we did not see any obvious adverse effects in this (nonsurvival) study; to our knowledge, little is known about whether the pharmacodynamic and pharmacokinetic parameters for tranexamic acid are similar between pigs and humans.
The neurophysiological recording setup proved well‐suited for obtaining accurate sensory evoked potentials directly from each nerve root.
It would have been optimal to obtain recordings from all eight selected nerve roots simultaneously as opposed to two separate rounds of recordings with four nerve roots each. However, we were limited by restrictions on the four‐channel EMG equipment available for animal experimentation. Also, the recording needles were placed in delicate tissue and had to be kept carefully in place for the time‐consuming recording sessions, which involved frequent movements of the experimental animal for skin needle repositioning. Keeping the needles correctly and stably positioned in the nerves proved increasingly technically challenging due to the number of roots that were simultaneously involved.
For stimulation of the skin fields, needle electrodes were selected to ensure a better impedance and thus a more effective (lower intensity) stimulation of the skin. The needles were placed as superficially as possible intra‐epidermally, and the intensity of the current was held under the threshold for a motor response of the underlying muscle tissue. To validate the method, we performed pilot examinations to measure the distance between the recording electrodes and the site of stimulation (e.g. latency 6 ms, distance 30 cm yielding a conduction velocity of 50 m s−1, which could agree with the range of conductions velocity of Aβ fibers).
We chose to create the skin fields based on anatomical landmarks rather than attempting to create estimated dermatomes directly comparable to those known from the classical medical literature.
First, we deemed it unlikely that we would manage to create probable dermatomes by transferring knowledge of their course in humans directly to pigs. Besides other embryological differences, one important difference between the human and the porcine neuroanatomy is that, in pigs, the spinal cord stretches all the way along the spinal column to the sacrum, rather than terminating around L1 to form a cauda equina as in humans. This, in our opinion, makes it even more futile to try to guess the course of dermatomes based on a human equivalent.
Secondly, we wanted to make skin fields that would allow us to reproduce them consistently from one animal to the other.
Experimental animals approximately 18 months old were chosen because the last epiphyseal plates close at this age (Bollen et al. 2010).
The downside of our approach is, of course, that the created skin fields are likely to overlap more than one dermatome.
Obviously, the larger the number of fields, the more accurate the approximated dermatome map. The number of recording fields in this study represents a compromise between what would be desirable and what proved practically possible. Each experimental session lasted around 12 h from induction of anesthesia to sacrifice of the animal; considering the extensive surgery performed, we did not deem it feasible to extend this period longer.
It should be noted that we carefully placed the needles in the geometrical center of each field to alleviate the interindividual variation that would be the result of random recordings of the various field parts.
The predefined skin fields were innervated by more than one nerve root, most likely because the roots span two or more dermatomes. Another explanation could be related to the nature of the dermatomes.
The classic perception of the dermatome map is one of well‐defined segmental areas innervated by one root only. From electrophysiological studies performed in humans, however, a large degree of interindividual variation is known to exist, concerning segmental innervations both of the musculature and of the skin (Liguori et al. 1992; Owen et al. 1993). It is also of note that there is a certain variation in the published dermatome maps (for review, see Lee et al. 2008). Thus, a dermatome should be considered a mere representation of areas with a dominant nerve root in the average population.
We do not consider the variation found in this series of experiments to exceed that found in equivalent human studies. With the reservations listed above, the data obtained can form the basis for an approximated dermatome map of the minipig hind body.
The approach to the data analysis described in the Statistical analysis section was chosen to avoid any systematic errors that could arise from any interindividual differences in sensory thresholds and a resulting difference in amplitude size. The pig(s) with the generally highest evoked potential amplitudes would then risk contributing disproportionately more to the general picture.
To eliminate this, we calculated all amplitudes as relative values (with the amplitude for each recording as a proportion of the mean amplitude of the individual pigs) before calculating the average value.
Surprisingly, this did not greatly change the overall picture compared with a simple averaging of all pigs irrespective of their mean amplitudes (data not shown). We carried out the same analysis, only with the amplitude of each potential as a relative value compared with the maximum amplitude achieved from each individual animal; this also yielded almost identical results (data not shown).
We considered expanding the analysis to include the neighboring fields, so that the results from each root of the fields surrounding the field in question would add to the result with a predefined weight; a spillover effect to compensate for the fact that the porcine dermatomes likely overlaps the skin fields. However, we chose not to make such an analysis because we deemed it too speculative and thus would not add any true value to the results presented.
To see whether the difference in amplitudes of the evoked potentials between each field could be a simple question of the amplitude used to generate the response (in other words, if we only got a high signal because we overstimulated), we compared each stimulation amplitude with the evoked amplitude and found no correlation. Thus, we conclude that our approach to selection of stimulation amplitude is viable.
As was the case for the sensory innervation of the skin fields, all muscles were innervated by more than one nerve root.
According to the extensive reference works by Nickel et al. (1986, 1992), the seven examined muscles are innervated as displayed in Table 1. The corresponding roots for each nerve listed in the table contain the information we were able to find based on information from Nickel et al. and from Getty's reference work (Sisson et al. 1975). The last column lists the results from our study for comparison.
Table 1.
Muscle innervation
| Muscle | Innervation | Spinal nerve root contributions to nerve | |
|---|---|---|---|
| Dissection studies | Our data | ||
| m. biceps femoris | n. glutaeus caudalis/n. tibialis | L5, L6, S1, S2 | L5, L6, S1 |
| m. peroneus tertius | n. fibularis | L5, L6, S1, S2* | L5, L6 |
| m. semitendinosus | n. glutaeus caudalis/n. tibialis | L5, L6, S1, S2 | L5, L6, S1, S2, (S3)** |
| m. tensor fascia latae | n. glutaeus cranialis | L5, L6, S1 | L4, L5, L6 |
| m. gluteus medius | n. glutaeus cranialis | L5, L6, S1 | L5, L6, S1, S2 |
| m. soleus | n. tibialis | L5, L6, S1, S2* | L5, L6, S1, S2 |
| m. coccygeus | nn. rectales caudales | S4 with contribution from S3 | S3*** |
The table shows the seven examined muscles. For each muscle, the nervous innervation is displayed along with the nerve roots attributed to each nerve according to the available literature. The last column shows our data with the dominant root in bold.
*Information only available for n. ischiadicus.
**Only one pig had a response from S3.
***S4 not tested.
In this series of experiments, we did not aim to map the motor innervation beyond the seven included muscles. The main focus of the study was to map the sensory innervation, and the muscle experiments served mostly as a pilot study for possible future experiments. A dedicated study on the motor innervation would require inclusion of several more muscle groups, but we showed that the experimental setup is entirely applicable for this purpose.
A 1991 report on a clinical study on human subjects studied motor responses of specific muscles of the lower limb after stimulation of spinal ventral roots (Phillips & Park, 1991). It is of note that the authors also found that the investigated muscles were innervated by several neighboring spinal roots.
The Göttingen minipig is steadily gaining recognition as an affordable large‐animal model well‐suited for neurological and neurosurgical research. Also, the animosity in the public opinion toward large‐animal research (particularly primates) seems less pronounced when pigs are concerned.
We were able to contribute to the increasing knowledge of the minipig nervous system, and our results can support the use of the Göttingen minipig as a model for experiments in medullar disorders and spinal cord‐related treatment modalities.
Disclosure of interests
The authors declare no conflict of interests.
Authorship
Kaare Meier (Principal investigator) designed the study and prepared the first draft of the paper. Erisela Qerama performed all neurophysiological recordings, contributed to the study design, and prepared most of the neurophysiology section of the paper. Kåre Schmidt Ettrup contributed to the animal experiments and to the preparation of the paper. Andreas Nørgaard Glud performed the graphical representation of the results and contributed to the preparation of the paper. Aage Kristian Olsen Alstrup acted as veterinary advisor, contributed to the animal experiments, and contributed to the preparation of the paper. Jens Christian Hedemann Sørensen performed the neurosurgical procedure, contributed to the study design and idea, and contributed to the preparation of the paper. All authors participated in the practical and theoretical work. All authors revised, contributed to, and approved the final manuscript.
Acknowledgements
The study was supported by the Danish Research Council and Department of Clinical Medicine, Aarhus University. The authors wish profoundly to thank the following individuals: for technical assistance: Lise Moberg Fitting, Center for Experimental Neuroscience, Aarhus University Hospital; for scientific advice: Prof. Anders Fuglsang‐Frederiksen, Department of Neurophysiology, and Prof. Troels Staehelin Jensen, Danish Pain Research Center, Aarhus University Hospital; for graphical representation of the results: Dr. Benjamin Svejgaard, Department of Biomedicine, Aarhus University, and Katrin Svabo Bech, MA; for statistical assistance: Prof. Michael Væth, Department of Public Health, Institute of Biostatistics, Aarhus University, and Dr. Ioanna Milidou, Perinatal Epidemiology Research Unit, Aarhus University Hospital; and for proof‐reading: Anne Sofie Møller Andersen, Department of Clinical Medicine, Aarhus University.
Appendix 1. Skin field description
The description details the division of the Göttingen minipig lower body into 36 predefined, consistently reproducible fields created using anatomical landmarks easily identifiable in the minipig surface anatomy.
Step 1: The following points were marked on the skin:
A: Anterior part of the hock joint
B: Posterior extremity of the hock joint
C: Anterior extremity of the femuro‐tibial articulation line
D: Deepest point of the popliteal fossa
E: Tuber coxae
F: Posterior midpoint of line drawn between both tubera coxae (L6)
G: Immediately anterior to vulva
H: Intersection of the inguinal grooves
I: Deepest point in the anterior inguinal fossa
J: Tuber ischiadicum
K: Base of digit 2
L: Base of digit 3
M: Base of digit 4
N: Base of digit 5
O: Posterior base of the tail
P: Approximately one vertebral height rostral to F
Q: Approximately two vertebral heights rostral to F
R: Approximately three vertebral heights rostral to F
S: Approximate termination of circumferential line from F to thoraco‐abdominal border
T: Approximate termination of circumferential line from P to thoraco‐abdominal border
U: Approximate termination of circumferential line from Q to thoraco‐abdominal border
V: Approximate termination of circumferential line from R to thoraco‐abdominal border.
Step 2: Based on the above, the following points were marked:
a: 1/3 of the distance from A to B (medial side)
b: 1/3 of the distance from A to B (lateral side)
c: 1/3 of the distance from B to A (medial side)
d: 1/3 of the distance from B to A (lateral side)
e: 1/3 of the distance from C to D (medial side)
f: 1/3 of the distance from C to D (lateral side)
g: 1/3 of the distance from D to C (medial side)
h: 1/3 of the distance from D to C (lateral side)
i: 1/3 of the distance from I to H
j: 1/3 of the distance from H to I
k: Midpoint of the distance from H to G
l: Midpoint of the distance from G to J
m: Midpoint of the distance from J to h
n: Midpoint of the distance from h to f
o: 1/3 of the distance from O to F
p: 1/3 of the distance from F to O
q: 1/3 of the distance from o to n
r: 1/3 of the distance from n to o
s: 1/3 of the distance from p to f
t: 1/3 of the distance from f to p
u: Midpoint of the distance from F to S
v: Midpoint of the distance from P to T
x: Midpoint of the distance from Q to U
y: Midpoint of the distance from R to V
z: Cross‐section of the line F to v and the continuation of the line from s through E
ae: Midpoint of the distance from a to e
bf: Midpoint of the distance from b to f
cg: Midpoint of the distance from c to g
dh: Midpoint of the distance from d to h
mk: Midpoint of the distance from m to k.
Step 3: Based on the above, the following lines were drawn:
A‐a‐c‐B‐d‐b‐A
K‐c
L‐a
M‐b
N‐d
thus dividing the foot into four parts: (1), (2), (3), and (4).
C‐e‐g‐D‐h‐f‐C
a‐ae‐e
c‐cg‐g
b‐bf‐b
d‐dh‐h
ae‐cg‐bf‐dh
thus dividing the crus into eight parts: (5), (6), (7), (8), (9), (10), (11), and (12).
I‐i‐j‐H
I‐C
H‐D
e‐i
g‐j
thus dividing the medial femur into three parts: (13), (14), and (15).
H‐k‐G
G‐l‐J
J‐m‐h
h‐D‐H
thus dividing the posterior femur into four parts: groin (16) and (17) and posterolateral (18) and (19).
E‐s‐q‐J
J‐m‐h
h‐n‐f‐C
C‐u‐E
m‐r‐t‐u
thus dividing the lateral femur into six parts: (20), (21), (22), (23), (24), and (25).
E‐u‐C
C‐S
S‐v‐z
z‐E
thus dividing the posterior flank into two parts: (26) and (27).
J‐O
O‐o‐p‐F
F‐z
z‐E‐s‐q‐J
thus dividing the sacral region into three parts: (28), (29), and (30).
S‐v‐z‐F
F‐P‐Q‐R
R‐y‐V
V‐U‐T‐S
v‐w‐x‐y
‐ thus dividing the anterior flank into six parts: (31), (32), (33), (34), (35), and (36).
Step 4: The resulting skin fields were marked and named:
1 Foot, anterior
2 Foot, medial
3 Foot, posterior
4 Foot, lateral
5 Crus, anterior, distal
6 Crus, medial, distal
7 Crus, posterior, distal
8 Crus, lateral, distal
9 Crus, anterior, proximal
10 Crus, medial, proximal
11 Crus, posterior, proximal
12 Crus, lateral, proximal
13 Femur, medial, rostral
14 Femur, medial, median
15 Femur, medial, caudal
16 Groin, rostral
17 Groin, caudal
18 Femur, posterior, distal
19 Femur, posterior, proximal
20 Femur, lateral, caudal, distal
21 Femur, lateral, caudal, proximal
22 Femur, lateral, median, distal
23 Femur, lateral, median, proximal
24 Femur, lateral, rostral, distal
25 Femur, lateral, rostral, proximal
26 Hip, distal
27 Hip, proximal
28 Sacral, caudal
29 Sacral, median
30 Sacral, rostral
31 Lumbar, caudal, distal (L5)
32 Lumbar, median, distal (L4)
33 Lumbar, rostral, distal (L3)
34 Lumbar, caudal, proximal (L5)
35 Lumbar, median, proximal (L4)
36 Lumbar, rostral, proximal (L3).
References
- Bollen PJA, Hansen AK, Alstrup AKA, et al. (2010) The Laboratory Swine, 2nd edn Boca Raton: CRC Press. [Google Scholar]
- Bosa YM, Getty R (1969) Somatic and autonomic nerves of the lumbar, sacral and coccygeal regions of the domestic pig (Sus scrofa domesticus). Iowa State J Sci 44, 45–77. [Google Scholar]
- Ettrup KS, Glud AN, Orlowski D, et al. (2011) Basic surgical techniques in the Göttingen minipig: intubation, bladder catheterization, femoral vessel catheterization, and transcardial perfusion. J Vis Exp (52), 2652. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Foditsch EE, Zimmermann R (2016) Development of a CT‐guided standard approach for tined lead implantation at the sacral nerve root S3 in minipigs for chronic neuromodulation. Res Rep Urol 8, 169–173. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Guiho T, Coste CA, Delleci C, et al. (2016) An intermediate animal model of spinal cord stimulation. Eur J Transl Myol 26, 6034. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Gutierrez K, Dicks N, Glanzner WG, et al. (2015) Efficacy of the porcine species in biomedical research. Front Genet 6, 293. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Hachmann JT, Jeong JH, Grahn PJ, et al. (2013) Large animal model for development of functional restoration paradigms using epidural and intraspinal stimulation. PLoS ONE 8, e81443. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Hort‐Legrand C, Noah L, Meriguet E, et al. (2006) Motor and sensory nerve conduction velocities in Yucatan minipigs. Lab Anim 40, 53–57. [DOI] [PubMed] [Google Scholar]
- Howard‐Quijano K, Takamiya T, Dale EA, et al. (2017) Spinal cord stimulation reduces ventricular arrhythmias during acute ischemia by attenuation of regional myocardial excitability. Am J Physiol Heart Circ Physiol 313, H421–H431. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Kowalski KE, Kowalski T, DiMarco AF (2016) Safety assessment of epidural wire electrodes for cough production in a chronic pig model of spinal cord injury. J Neurosci Methods 268, 98–105. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Lee MW, McPhee RW, Stringer MD (2008) An evidence‐based approach to human dermatomes. Clin Anat 21, 363–373. [DOI] [PubMed] [Google Scholar]
- Liguori R, Krarup C, Trojaborg W (1992) Determination of the segmental sensory and motor innervation of the lumbosacral spinal nerves. An electrophysiological study. Brain 115(Pt 3), 915–934. [DOI] [PubMed] [Google Scholar]
- Nickel R, Schummer A, Seiferle E, et al. (1986) The Locomotor System of the Domestic Animals, 5th edn, Vol. 1 Berlin: Verlag Paul Parey. [Google Scholar]
- Nickel R, Schummer A, Seiferle E (1992) Lehrbuch der Anatomie der Haustiere In: Nervensystem, Sinnesorgane, Endokrine Drüsen, 3rd edn, Vol. 4 (ed.Böhme G.), Berlin: Verlag Paul Parey. [Google Scholar]
- Odenstedt J, Linderoth B, Bergfeldt L, et al. (2011) Spinal cord stimulation effects on myocardial ischemia, infarct size, ventricular arrhythmia, and noninvasive electrophysiology in a porcine ischemia‐reperfusion model. Heart Rhythm 8, 892–898. [DOI] [PubMed] [Google Scholar]
- Owen JH, Bridwell KH, Lenke LG (1993) Innervation pattern of dorsal roots and their effects on the specificity of dermatomal somatosensory evoked potentials. Spine (Phila Pa 1976) 18, 748–754. [DOI] [PubMed] [Google Scholar]
- Phillips LH 2nd, Park TS (1991) Electrophysiologic mapping of the segmental anatomy of the muscles of the lower extremity. Muscle Nerve 14, 1213–1218. [DOI] [PubMed] [Google Scholar]
- Sack WO (1982) Essentials of Pig Anatomy/Horowitz/Kramer Atlas of Musculoskeletal Anatomy of the Pig. New York: Veterinary Textbooks. [Google Scholar]
- Sisson S, Grossman JD, Getty R (1975). Sisson and Grossman's the Anatomy of the Domestic Animals, 5th edn, Vol. 2 (ed. Getty R.), Philadelphia: Saunders. [Google Scholar]
- Solis LR, Twist E, Seres P, et al. (2013) Prevention of deep tissue injury through muscle contractions induced by intermittent electrical stimulation after spinal cord injury in pigs. J Appl Physiol (1985) 114, 286–296. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Sondeen JL, Hanson MA, Prince MD, et al. (2016) Double‐blinded, placebo‐controlled study of early tranexamic acid treatment in swine uncontrolled hemorrhage model. J Trauma Acute Care Surg 80, 81–88. [DOI] [PubMed] [Google Scholar]
- Sorensen JC, Nielsen MS, Rosendal F, et al. (2011) Development of neuromodulation treatments in a large animal model – do neurosurgeons dream of electric pigs? Prog Brain Res 194, 97–103. [DOI] [PubMed] [Google Scholar]
- Zentai C, van der Meijden PE, Braunschweig T, et al. (2016) Hemostatic therapy using tranexamic acid and coagulation factor concentrates in a model of traumatic liver injury. Anesth Analg 123, 38–48. [DOI] [PubMed] [Google Scholar]