Jugular Vein Catheter Design and Cocaine Self-Administration Using Mice: A Comprehensive Method

Intravenous self-administration (IVSA) is a behavioral method of voluntary drug intake in animal models which is used to study the reinforcing effects of drugs of abuse. It is considered to have greater face validity in the study of substance use and abuse than other assays, and thus, allows for valuable insight into the neurobiological basis of addiction, and the development of substance abuse disorders. The technique typically involves surgically inserting a catheter into the jugular vein, which enables the infusion of drug solution after the performance of a desired operant behavior. Two nose- poke ports or levers are offered as manipulanda and are randomly assigned as active (reinforced) or inactive (non-reinforced) to allow for the examination of discrimination in the assessment of learning. Here, we describe our methodological approach to this assay in a mouse model, including construction and surgical implantation of a jugular vein catheter, set up of operant chambers, and considerations during each phase of the operant task.


INTRODUCTION
Drug intravenous self-administration (IVSA) is a behavioral test that has become widely used and accepted in the study of the reinforcing effects of drugs of abuse using animal models, including rats, non-human primates, cats, and mice. It offers valuable insight into the neurobiological basis of addiction by mimicking real-world behaviors that are associated with recurrent drug use during the development and maintenance of substance abuse disorders. Drug IVSA involves surgically inserting an intravenous catheter, typically into the jugular vein, and uses the principles of operant conditioning, such that the subject learns to perform a desired behavior to receive drug delivery (Panlilio and Goldberg, 2007). Typically, the drug IVSA test is conducted in operant chambers that offer two identical manipulanda-items in the conditioning chamber that can be manipulated physically-such as nose poke ports or levers. Manipulanda are assigned as active (reinforced) and inactive (non-reinforced) to allow both for the analysis of port discrimination in the assessment of learning, as well as for general activity levels. A reinforcement schedule allows the experimenter to modify task requirements for the type of research being conducted (Platt and Rowlett, 2012).
Since IVSA allows for volitional control of intake, it is considered to have higher face validity for addictive behavior than other types of tests, and has been used, for example, to verify that nicotine mediates reinforcement via activity at nicotinic acetylcholine receptors (Fowler and Kenny, 2011). An alternative method is oral self-administration, which can be conducted in both two-bottle

Advantages and disadvantages of intravenous vs. oral drug self-administration models
Advantages of IVSA • Onset of drug action for IVSA is more rapid than oral SA, and thus, better mimics certain types of human drug abuse. • IVSA allows more assurance, as well as accurate determination of the volume, of substance delivery. • IVSA allows higher bioavailability of substances, bypassing the need for absorption through the digestive tract, avoiding digestion-related substance degradation, and avoiding first-pass metabolism by the liver. • IVSA eliminates intake alterations or avoidance related to taste.

Disadvantages of IVSA
• IVSA is less convenient and safe than oral SA, necessitating major surgery and introducing opportunities for infection. • IVSA is less economical, requiring single-use catheters and surgical materials. • IVSA has lower throughput than oral SA due to loss of catheter patency and procedure-related mortality in a subset of animals. • IVSA requires the continued presence of invasive materials in/on the body and catheters must be maintained by regular flushing. • IVSA always requires some level of active training/learning, whereas some forms of oral self-administration (e.g., two-bottle choice) are minimally dependent on such processes.
choice paradigms, as well as operant paradigms; however, there are trade-offs to consider for the two methods (refer to Table 1). Advantages of the IVSA method include allowing the experimenter to remove possible confounding variables and accurate quantification of the amount of drug consumed. Along with increased accuracy, the IVSA allows for a rapid increase in drug levels in the blood and brain (Kmiotek et al., 2012). The IVSA is also the gold-standard test for assessing the addictive propensity of novel compounds, and the significance of prior drug exposure can be investigated. In addition, infusions can be paired with discriminative cues to allow for later evaluation of drug-seeking behaviors following re-exposure to a reinforcerpaired cue (Thomsen and Caine, 2005), which is often used as a model of relapse (Galaj et al., 2016). Abstinence, or time without access to a drug, causes a deprivation effect and can be used to assess drug-seeking (when no drug is available) or changes in administration upon renewed drug access. For example, "incubation" of craving has been demonstrated for cocaine, where cocaine-seeking escalates between 1 and 45 days of abstinence following prior repeated exposure. Using the drug IV self-administration assay, many factors, both intrinsic and extrinsic, have been shown to influence voluntary drug-taking. Intrinsic factors, such as exploratory and risk-taking behaviors in mice, are associated with an increased propensity to self-administer cocaine (Dickson et al., 2016). Sex is also an important factor, since female rats self-administer more nicotine, for example, than male rats (Galankin et al., 2010;Flores et al., 2019). Extrinsic factors, such as social learning and environment, also influence the tendency of an animal to self-administer drugs. A study in which rats had to drink a saccharin solution to get infusions of nicotine

Comparison of mice and rats for use in IVSA
Basic study logistics • Using mouse models allows leverage of a vast array of genetic models.
• Mice typically have cheaper housing facility costs.
• Smaller testing equipment for mice allows more to fit into a lab space.
• Mice may be less destructive to some materials and equipment.
Surgical outcomes • Rats have larger anatomical structures, which facilitate surgery and aide catheter patency, but the smaller anatomy of mice can be offset by thoughtful catheter choices and patient, observant surgical practice. • Greater irritation and rupture of skin around the catheter base in mice can be significantly reduced by using catheters with highly pliable (e.g., monofilament polypropylene) surgical mesh. • Rats may have lower surgery-associated mortality by conventional methods, but survival is improved in mice by using sevoflurane, which speeds recovery from anesthesia. Special attention to humidity and hydration also improves outcomes in mice (Thomsen and Caine, 2005).
Task feasibility • Rats are credited with ability to perform more complex tasks than mice; however, we find mice capable of tasks sometimes labeled as "too difficult" for them (e.g., reinstatement of cocaine IVSA, increased schedules of reinforcement). Being natural prey animals, fear management (e.g., preparatory handling sessions prior to procedures, calm experimenters, test room acclimation) is essential in mice.
revealed that interaction with another rat drinking the solution increased the likelihood of stable drug-taking behavior (Wang et al., 2016). There are mixed findings on the impact of diet, with studies showing that high-energy diets either do not change (Bruggeman et al., 2011) or suppress cocaine reinforcement (Wellman et al., 2007). The IVSA has been shown to be reduced following exercise, with rats reducing methamphetamine intake after just 1 day of access to an exercise wheel (Aarde et al., 2015). The acquisition of the drug IVSA in rats may also be elevated by higher ambient temperature (Aarde et al., 2017). In comparison to basic laboratory animal housing, enrichments, such as toys, obstacles, and running wheels decrease the drug self-administration (Ewing and Ranaldi, 2018). These findings have broad implications for the types of questions that can be answered using this important behavioral technique. Rodents are often used in IVSA studies, but there exists a historical preference for using rats over mice, likely due to feasibility (refer to Table 2 for a comparison of the two species). In rats, the earliest veinous catheters for drug delivery emerged by the early 1960s (Slusher and Browning, 1961;Weeks, 1962), while the first mouse catheters were described nearly 20 years later (Barr et al., 1979). However, self-administration of substances in mice first relied on tail-vein injection without catheters (Criswell, 1982;Criswell and Ridings, 1983), and mouse self-administration using chronic indwelling catheters was only described later (Carney et al., 1991;Grahame et al., 1995;Deroche et al., 1997). However, with the development of new and smaller materials and   model-related motives for using mice to test the role of genetics in substance abuse, more labs are attempting mouse IVSA. In the current methodological description, founded originally on methods developed by Thomsen and Caine (2005), we focus on cocaine IVSA using a mouse to model and measure the addiction-related behavior. Our primary goals in this manuscript are to increase the transparency of methodological choices in mouse IVSA, and likewise, increase the accessibility of the technique. We describe the surgical procedure we use and demonstrate and explain the importance of different phases of the behavioral test, including acquisition, extinction, measurement of a dose-response curve, and increasing the cost. We also discuss other methodological approaches that may be selected for different purposes.  Table 3. Details and images of an assembled catheter can be seen in Figure 1, and tips for troubleshooting surgery and behavioral testing can be found in Table 4. 2.1.1.2 Cap 1 mL syringe with blunt-tipped needle and attach tubing to the end as shown in Figure 1A.

MATERIALS AND EQUIPMENT
Tubing-attached syringes will be used for checking catheter patency, surgical drug delivery, and catheter flushing, so it may be useful to prepare multiple syringes. Magnetic twist ties can be used to elevate and help secure the anesthesia nose cone.
The mouse jugular vein cannot be located.
The pectoral muscle lays like a pink blanket tucking in the jugular vein, which peeks out from the top. Once the skin incision is made midway between sternum and the ear, use forceps to gently slide top layers of tissue toward the animal's left side, which often reveals the dark red vein. If not successful, use forceps to break the overlying clear fascia, then repeat above. If not successful, puncture the visible layer of tissue discretely and continue to pull aside top layers as you look for the vein.
While clearing tissue from the jugular vein, the vein tears.
Immediately staunch the bleeding with firm but gentle pressure to the vein or by lifting the surgical bar. If the bar is not in place, try to make a small space to insert it while continuing to control any bleeding. It is possible to place a catheter in a torn vein if it is not severed.
To avoid tearing the vein, remove fatty tissue by gripping it away from the vein and slowly pulling in a motion parallel to the vein (not perpendicular). Do not grip the vein directly or pull the clear surface of the vein. Opaque tissue fibers that remain close to the vein can be separated and cut against the surgical bar.
After catheter placement in the vein, blood cannot be pulled back using the syringe unless the catheter is less than maximally inserted.
When testing catheter insertion, occasionally it is necessary to partly slide the catheter out of the vein (without removing it) to allow blood flow, after which the catheter can be pushed back into place. If after 2-3 tries, blood is not flowing easily, then the 1.2 cm length of catheter (from anchor to tip) is too long. Cutting the catheter to 1.0-1.1 cm length may help. With experience, shortening can be done during surgery.
Death occurs when using catheter sled.
Vulnerability to this issue differs by mouse size and background strain. The sled should be inserted minimally and not more than 5 mm. A stack of gauze near the mouse's head should be used to keep the winged end of the sled propped at the level of the vein or slightly above.
The catheter cannot be fully pushed in or pushes back out. AND/OR No blood is seen at pullback and pulling the catheter out halfway does not fix this.
The catheter is inserted between the vein wall and a sheathe that surrounds the vein. Double check that the vein cut resulted in bleeding. If not, or if there is a tiny amount of blood, make a slightly larger cut at the same position. If bleeding indicates the cut is already adequate, shifting the catheter (or sled) introduction point up or down the vein length from the perceived cut can help.
Attempts at vein entry are not successful.
Minimize manipulation of the vein at every opportunity, as it tends to shrink the vein. Target the catheter (or sled) to enter the vein slightly higher or lower than the perceived cut. Check whether bleeding indicates a sufficient cut and consider improving the cut.
If the sled has not been used, try it. If too much time has passed, blood may have clotted. Use rinsing syringe of saline to clear and moisten vein. If saline does not help, a new cut may be required.
Surgery time and mortality need to be reduced.
Improve technique by inserting the catheter without a sled. While cutting the vein, look closely for the exit point of blood, then move swiftly to introduce the catheter tip firmly at that site, using a lateral and downward motion (toward the mouse's vein and feet).
If not successful, a sled can still be used.
Blood was flowing with syringe pullback, but now there is no blood flow with pullback.
Vein suture ties may be too tight. Check blood flow just after securing the ties. If flow is diminished, minimally loosen the ties one-at-a-time and recheck, repeating until flow is restored. If blood flow was confirmed after vein sutures were tied but stops after the mouse is supine, the tubing may be constrained. From the back incision, seat the catheter base and slightly rotate it left or right to adjust the direction of the tubing as it exits the base, while checking blood flow. If blood remained visible in the tubing after a blood flow check, it may have coagulated. Ideally, pushing a tiny volume of saline into the line should move blood easily. If not, gently alternating between pushing and pulling the syringe plunger may loosen the clotting. Avoid pushing much saline into the animal as it can result in death. If blood flow cannot be confirmed before ending surgery, the back and/or chest incisions will need to be reopened, and tubing direction, vein sutures, and catheter entry reexamined, checking for blood flow until it has returned.
Mice die or become ill during recovery.
Give post-surgery and recovery administrations of USP-grade saline (0.5-1.0 mL per mouse, s.c., 1-3 times/day). Separate mice recovering from anesthesia from cage mates not anesthetized for at least 24 h. If mice become ill despite antibiotic use, ask your facility vet to test for presence of harmful, systemic bacteria and advice on antibiotic choice.

During operant conditioning
Mice are slow to acquire task. Ensure mice are not sitting in chambers for long after sessions have ended. Some or all these interventions can be used: vanilla flavored Ensure ® around/on active manipulandum; 1-3 non-contingent (priming) drug infusions (depending on concentration) either at session beginning or ∼15 min into session; increased drug concentration for 1 session; 1-3 overnight sessions (lengthen session and use timeout periods to minimize overdose risk). Check catheter is patent. If using levers, consider changing to nose poke ports.
Should timeout periods or response maximums be used?
Timeout periods following drug delivery and limitations on the maximum number of responses allowed in a session may restrict the normal range of drug-taking behavior. Consider only using timeouts >3 s for very high doses or when switching from a long period of low-moderate dose availability to a much higher dose.

(Continued)
Frontiers in Behavioral Neuroscience | www.frontiersin.org Mice are slow to extinguish. Sessions may be lengthened to 5-6 h temporarily. If extinction sessions include the discriminative cue (reinstatement not planned), then response requirements can be increased temporarily (i.e., FR3 or above).
Catheter malfunction or loss of catheter patency during experiment.
Remove the catheter, implant a new one into the jugular vein on the mouse's left side, and allow recovery. If the study uses criteria to determine completion of each phase, run each phase as before, repeating any that were already run. Data from phases obtained originally with a patent catheter should be kept, but use "recath" data for any phases not completed with a patent catheter in the original run. Note: this approach may not be appropriate for all study designs. Before beginning, plan how to appropriately combine each animal's original and post-"recath" data depending on study goals and design. (D) Closer view of tether with connector matching the catheter guide cannula threading. Tether length may need to be shortened and weight position changed on the counterbalanced arm so that the tether has an ideal amount of slack; it should not pull down the catheter or impede the mouse's ability to move or access all areas of the chamber floor.

Prepare Catheter Base With Custom Tubing
2.1.2.1 Slide custom cannula over the longer (10 mm) guide cannula end, which will exit the bottom of the catheter base (refer to Figure 1B Figure 1C). Dip a cotton-tipped applicator in mold release and apply to the inside of each receptacle in the catheter base mold.    Table 5) and vary the infusion time according to mouse weight. One option is described in Figure 3.      In some studies, the house light remains on as long as the program is in session. Alternatively, it may be more specifically used for the signal availability of the drug, remaining on outside of drug delivery and timeout periods.

METHODS
The animal study was reviewed and approved by Texas A&M University Institutional Animal Care and Use Committee. 3.1.4.2 With a hair trimmer nearby, anesthetize the mouse in the induction chamber (7% of sevoflurane in oxygen at ∼500 mL/min rate), always maintaining the supervision of the mouse. 3.1.4.3 When breathing is slowed and even, remove the mouse, and shave from just behind the ears to mid-back in a large square. Repeat anesthesia and shaving, if necessary.

Jugular Vein Catheter Implantation Surgery
3.1.5. Anesthesia Maintenance and Pre-operative Care 3.1.5.1 To begin surgery, induce anesthesia again, then rapidly move the mouse to the face mask or nose cone delivery device and adjust the anesthesia settings as appropriate for the maintenance of anesthesia (2-3% sevoflurane; adjust the flow rate according to vaporizer manufacturer instructions).
Monitor the mouse throughout the surgery. If there is gasping, adjust the dial to <2%, and if there is shallow breathing, adjust the dial toward 3%. 3.1.5.2 Next, apply eye lubricating ointment to completely cover each eye. Clean incision sites with an alcohol pad or betadine and allow to air dry. Using an insulin syringe, administer ketoprofen (0.05 mL; subcutaneous) into the lower back.

Position and Prepare Catheter
3.1.6.1 According to your animal protocol, confirm that the mouse is sufficiently anesthetized. Pull up the skin on the back of the neck using forceps, and at the midline, just behind the base of the ears, create a tiny hole with small surgical scissors. Insert the scissors and cut a ∼2 cm incision along the midline of the midscapular region toward the tail (Figure 4A). 3.1.6.2 To make space for the catheter base under the skin, insert the hemostats ∼1.5 cm into the incision, making slight opening and closing motions with the tips just beneath the skin in a circular area around the incision. Then, using the same motions with the hemostats, tunnel under the skin over the mouse's right shoulder toward the front of the chest. 3.1.6.3 Turn the mouse to a supine position, moving the face mask or nose cone accordingly 10 . 3.1.6.4 Rake the hairs on the right side of the mouse's chest to make them stand straight up and trim the hair close to the skin using small scissors. To locate the vein, look closely for jugular vein movement. Using forceps and small scissors, make a small incision from the sternum toward the ear over the vein 11 . 3.1.6.5 Attach the tubing-attached saline syringe to the sterilized catheter to flush and fill it with sterile saline. Leaving them connected, place the syringe and catheter on the platform near the mouse's right shoulder. 3.1.6.6 With hemostats, tunnel through the front incision under the skin to the back incision, opening and closing the hemostats very slightly along the way. Very gently pinch the beveled end of the flushed catheter with the hemostats and pull it through the tunnel to the front of the animal (Figure 4B). Turn the hemostat handle away from you and set it down 10 Make sure that the breathing of the mouse remains stable; if gasping is observed, anesthetic percent and/or flow need to be adjusted downward. 11 Make sure this incision and the vein are moistened regularly with a sterile saline throughout the duration of the surgery using the prepared rinsing syringe. behind the right shoulder of the mouse. Support the pointed end on a stack of gauze or with the workstation "helping hands" so that the hemostats are pointed slightly upward and hold the catheter tip out of the way.  Figure 4C) and carefully slide the short surgical bar under the vein to keep it isolated ( Figure 4D). 3.1.7.5 If more substantial tissue strands are still present alongside the vein and are visibly distinguishable from it, use forceps to sever them before moving on. 3.1.7.6 Cut two ∼1 ′′ lengths of suture thread. Using forceps, place them under the vein, one at the lower end (below the bar) and one at the upper (above the bar). Turn each thread into a very loose knot around the vein by crossing the ends and pushing one end through the hole twice. Only pull the ends of each knot slightly, so they make relatively wide but secure loops around the vein ( Figure 4D) (Figure 4E) 16 . If successful, slowly push the blood back into the animal, using care not to introduce much saline.  13 With experience, the catheter may be inserted directly into the vein opening. Look closely for the exit point of blood from the vein when the cut is made will help. If the first few attempts are not successful, targeting the catheter to enter the vein slightly higher than the perceived cut can sometimes increase success. 14 There are sheaths around the jugular vein that can give the appearance of a successful catheter entry. If the catheter cannot be pushed in fully and produces no blood at pullback, this sheath is often the cause. Double check that the cut made to the vein by the artery scissors resulted in bleeding. 15 Inserting the sled more than a minimal amount causes rapid death, vulnerability to which differs by mouse size and background strain. 16 Occasionally, it is necessary to slide the catheter back out of the vein slightly (without removing it) to allow for blood flow, after which the catheter can be pushed back into place. 17 It is easy to overtighten at this step, so use the syringe to check the blood flow, as described above, afterward. If blood flow is diminished, minimally loosen the ties one at a time and recheck the flow.  Figure 4F) and stitch the anterior end of the incision closed, starting at the catheter, in the manner described previously.
3.1.11. Post-operative Care (Immediate) 3.1.11.1 Using the tubing-connected syringe with cefazolin/heparin solution, push fluid to the tip of the tube, then connect and flush the catheter (0.02-0.03 mL). 3.1.11.2 Inject the mouse subcutaneously on one side of the lower back using a warmed hydration syringe. Apply triple antibiotic cream to both incisions. 3.1.11.3 Holding the catheter base securely, push a plastic cap onto the guide portion of the catheter using a twisting motion, then cover it with a metal dust cap before turning off the anesthetic gas. 3.1.11.4 Move the mouse to the recovery cage. Once the mice are awake and mobile, they can be moved to a clean housing cage with food and water. 3.1.11.5 At this point, the mice can be co-housed with any original cage mates which are also anesthetized on the same day.

Between Surgeries
3.1.12.1 Rinse instruments with water, bead sterilize them, and soak in ethanol. Clean and re-sterilize the surgical area, then move tools to the platform to dry. Apply triple antibiotic cream to incisions once daily for at least 3 days, and consider at least one saline injection, as described below, for hydration. The mice that are fully recovered from anesthesia and recovering as expected from surgery (i.e., not having a hunched, unkempt appearance, dehydration, etc.) can be rehoused with all original cage mates. 3.1.13.1.3 Starting on Day 9: Catheters should be flushed at least 5 days per week with heparin-saline (0.03 µL). During self-administration, catheters should be flushed both before and after each session. On days with flushing and no testing, once daily is sufficient. After recovery, catheter patency must be checked throughout the study 18 , as well as any time failure is suspected, by flushing it with ketamine/midazolam (15 mg/mL/0.75 mg/mL) solution (0.03 mL). Loss of righting reflex should be observed within 3 s of infusion, and mice not meeting this criterion should be removed from the study or may undergo surgery for new catheter placement.
3.1.14. Post-operative Complications 3.1.14.1 Post-surgery dehydration is common and potentially lethal. Mice suspected of dehydration should be given 0.5-1.0 mL (s.c.) of warmed saline, as described above for surgery, between 1-3 times/day. An animal slow to recover or having a hunched appearance that is not helped by subcutaneous saline should be given ketoprofen daily (1 mg/mL; 0.05 mL/day) until they improve. Triple antibiotic cream should be applied to incisions that are red or oozing. For all these conditions, the facility vet should be consulted as needed. 18 Three to four patency check point times should be planned, including just before the first and just after the last sessions. 19 A major factor in session length is whether escalation in drug-taking will be studied over time, for which 4-6 h sessions have been shown to be required. 20 Food training may occlude differences in acquisition dose studies, and particularly when no extinction is performed for food training, leaves open the possibility that later responding is driven by food-motivated seeking. Additionally, drug-taking behavior under food-restriction may have limited translational relevance to human drug use.

Experimental Timeline
Experimental timelines for IVSA studies typically include multiple phases of testing, as mentioned above, which may vary based on the objectives of the experiment. For example, a reinstatement phase will provide insight into drug-seeking behaviors after withdrawal (De Vries et al., 1998), while a doseresponse phase will identify changes in sensitivity, tolerance, or hedonic set point (Schenk and Partridge, 1997;Ahmed and Koob, 1998). One possible timeline, including acquisition, extinction, dose-response, and increasing cost schedule phases, is provided as an example in Figure 5. In the following sections, we provide anticipated results (reprinted with permission from Huebschman et al., 2021), discuss advantages, limitations, potential pitfalls, and troubleshooting options for each of these phases. Unless stated otherwise, operant sessions in all phases were 3 h per day and ran 5 days/week. Sessions were terminated early if the reinforcer limit was reached (1.0 mg/kg/inf = 30 max; all lower doses = 100 max; 3.2 mg/kg/inf = 10 max with 10 min timeouts if > 2 reinforcers in < 10 min), and all reinforcers were followed by a 20 s time-out period during which additional reinforcers were unavailable.

Acquisition
Acquisition sessions began with the illumination of the active port cue light and a single 1.0 mg/kg infusion of cocaine, after which nose-pokes in the active port resulted in reinforcement (1.0 mg/kg/infusion), as well as cue light illumination, on an FR1 schedule of reinforcement. This dose and schedule were selected to induce stable drug-taking in a large percentage of animals. To limit the contributions of overtraining and variation in total drug intake during this phase, each mouse remained in acquisition only until a predefined set of learning criteria was reached and thus individual animals differed in the number of sessions during this phase. These criteria were defined as two consecutive sessions having ≥15 reinforcers with no more than 20% variation in reinforcers earned between those sessions, and at least 70% discrimination for the active port. If "over-learning" is not a concern for one comparison group, researchers may opt to run acquisition sessions for a set number of days rather than to predefined learning criteria. Average nose-pokes, excluding responses made during timeout periods, for sessions where >25% of animals remained in the acquisition phase are shown in Figure 6A. A two-way repeated measures (RMs) ANOVA comparing nose-poke behavior on each animal's first and last day of acquisition ( Figure 6B) reveals a significant interaction of session and port [F (1,11) = 26.47, p < 0.001]. Follow-up one-way RM ANOVAs show a simple main effect (SME) of port at the last [F (1,11) = 234, p < 0.001], but not the first, session, with mice making significantly more nose-pokes at the active port than the inactive. SMEs of session were also significant for both the active [F (1,11) = 13.3, p < 0.01] and the inactive [F (1,11) = 13.7, p < 0.01] ports, with mice significantly increasing and decreasing in the nose-poke behavior at each port, respectively.
Of the 18 mice included in the acquisition phase, 12 met acquisition criteria within 15 sessions and progressed to later phases of testing. The rate at which these mice met the criteria is shown in Figure 6C. The medium dose used during this phase is intended to maximize the acquisition, and we typically expect ∼70% of control animals to meet the criteria. If mice struggle to acquire the task, and it does not interfere with the experimental objectives, interventions as in section 3.2.4.2 can be made to facilitate learning. In the study shown here, overnight sessions were given to animals that had not yet met the criteria after 5 and 10 days.

Extinction
Extinction sessions were conducted in the same manner as acquisition, except that active nose-pokes had no consequence. Animals remained in the extinction phase until active port responding dropped to <30% of the average of the last 2 days of acquisition. Average nose-pokes for sessions where >25% of animals per group remained in the extinction phase are shown in Figure 7A. Two-Way RM ANOVA of nose-poke responses on the first and last day of extinction showed a significant main effect of the session [F (1,11) = 5.87, p < 0.05], with mice making fewer nose-pokes overall on the last day ( Figure 7B). Of the 12 mice that underwent extinction, 2 failed to meet the phase criteria within 15 sessions and were excluded from further analysis. The rate at which the animals met the extinction criteria is shown in Figure 7C.
In this study, active-port nose-pokes during extinction had no consequence. Depending on the objectives of the study, extinction sessions may be run such that active nose-pokes either do or do not trigger cue-light illumination. The former helps ensure that, during drug delivery phases, animals respond to the active port for the reinforcer itself and not to the associated cues. On the other hand, the latter is necessary to examine cue-induced reinstatement after the extinction phase, as it is the reintroduction of the cue-light that drives the reinstatement of drug-seeking behavior. Ideally, in either case, saline replaces drug infusion during extinction, a practice that also maintains any sound cues associated with pump operation. However, if necessary, infusions can be omitted. A small percentage of animals (∼15% in our hands) may take 30-45 days to meet the extinction criteria, so it may be beneficial to set a cutoff point at the onset of the study. In the example provided here, we used a cutoff of 15 sessions, and mice that took longer were omitted from the additional testing phases.

Dose-Response and Increasing Cost Schedule
After completing the extinction phase of testing, animals began re-acquisition sessions, which were identical to acquisition sessions and conducted until reinforcers earned in a single session returned to ≥15. All mice met these criteria within 1-2 sessions. For dose-response testing, a single concentration (0, 0.01, 0.032, 0.1, 0.32, 1.0, and 3.2 mg/kg/infusion, in 0.9% saline; see Table 5) was made available for each session. Doses were presented in sequential order, with the starting dose counterbalanced across groups using a Latin square design. Twoway RM ANOVA revealed a significant interaction of dose and port [F (2.11,12.69) = 9.64, p < 0.01], with follow up one-way RM ANOVA showing a significant effect of dose for the active port alone [F (2.08,16.63) = 14.9, p < 0.001] (Figure 8A). Post-hoc Bonferroni pairwise comparison indicates that responses on the active port for the 0.32 dose were significantly higher than those for the 0.01, 0.032, 1.0, and 3.2 mg/kg/infusion doses (p < 0.05, 0.05, 0.01, 0.01, respectively) and responses for the 1.0 dose were significantly greater than for the 3.2 dose (p < 0.001).
Following dose-response testing, mice returned to the acquisition dose (1.0 mg/kg/infusion) on an FR1, followed by an FR3, and then an FR5 schedule of reinforcement (two consecutive sessions each) ( Figure 8B). Three-way RM ANOVA revealed a significant interaction of cost schedule with port responses [F (2,8) = 9.42, p < 0.01], and follow up one-way RM ANOVA identified a significant effect of cost schedule for active port responses alone [F (2,18) = 5.71, p < 0.05]. Post-hoc Bonferroni pairwise comparison showed a significant difference in active port responses between FR1 and FR5 (p < 0.05), with mice making fewer responses at the higher schedule of reinforcement.
Inverted U-shaped dose-response curves, as seen here, are typical of drug self-administration tasks, as rodents adjust their responses to maintain targeted minimum satiety (Tsibulsky and Norman, 1999). Dose-response testing can provide valuable information, as left or right horizontal shifts in the dose-response curve indicate changes in sensitivity or tolerance (Schenk and Partridge, 1997). On the other hand, vertical shifts have been associated with altered hedonic set points after prior drug experience or escalating drug intake (Ahmed and Koob, 1998). Upward shifts are also observed in animals that display stronger responses on increasing cost schedules of reinforcement (Piazza et al., 2000), so the inclusion of both phases of testing allows for a more robust characterization of behavioral phenotypes.

DISCUSSION
Intravenous self-administration using operant conditioning techniques has proven a preferred way of studying substance abuse due to its relative resemblance to drug-taking behavior observed in humans. However, amongst rodent animal models, the rat has predominantly been used for IVSA, likely due to technical challenges associated with implanting and maintaining catheters in mice. However, as several labs have demonstrated, the IVSA can be performed in a mouse model and could be more accessible to labs requiring the method. In the Methods section, we describe our optimized approach for conducting cocaine IVSA in mice, including design details for a custom, back-mount catheter that can be prepared in-house, as well as detailed technical surgical advice tailored to jugular vein catheter implantation in this species. We show that these methods result in successful cocaine self-administration using data from the wildtype mouse group in our recently published manuscript, which can be consulted for additional details (Huebschman et al., 2021). In the 2021 study, we moved mice through the acquisition and extinction phases of cocaine IVSA based on when they met the criteria set for learning, then we performed tests for dose-response and increased cost. As expected, wildtype C57BL/6N mice acquired and extinguished the task quickly, showed an inverted u-shaped curve during dose-response testing, and earned fewer reinforcers as schedule requirements were increased from FR1 to FR5.
When we look more closely at our representative results, we see that wildtype mice show a clear preference over the course of acquisition for the active (cocaine + cue light) instead of the inactive (no effect) port. While running each phase to criteria prevents us from cleanly assessing learning over the same number of sessions, mice averaged ∼25 active port responses on the second day of meeting criteria requirements. Reflecting this level of intake, mice remaining in sessions six through 10 show, on average, from 20 to greater than 25 active port responses (per 3h session). These results resemble findings from others testing cocaine IVSA in previously naïve wildtype mice. For example, using the same acquisition dose and schedule, prior work shows wildtype mouse making ∼23 active port responses per 3-h session over the 2 days when criteria were met (Thomsen et al., 2009b). We note that mice shown in this study were required to meet (slightly more stringent) acquisition criteria within seven days, while mice shown in our study took up to 10 days. In another study, C57BL/6 wildtype mice at the same acquisition dose and schedule showed between 20 and 25 infusions/90-min session on days of 6-10 of acquisition, though they received three of these shorter sessions/day (van der Veen et al., 2008). C57BL/6J wildtype mice receiving the same 1.0 mg/kg/infusion dose and FR1 schedule responded for around 29 infusions per 3-h sessions at the point of meeting the acquisition criteria (Stoll et al., 2018).
During extinction, we required more stringent criteria (<30% of the averaged last 2 days of acquisition) than some other published studies [e.g., <80% (Thomsen et al., 2009b)]. As such, response levels are difficult to directly compare, but as expected, our mice average lower responses on the last day of extinction. For example, the average active port response number on the last session of our extinction phase was around five, while the above studies showed ∼13 (Thomsen et al., 2009b). Also, a study using the extinction criteria of ≤30% of acquisition level compared to ours very closely, showing about seven active port responses at the criteria (Stoll et al., 2018). The authors report around 16-18 sessions on average to extinction criteria for their control groups, while our mice met extinction criteria, on average, in 8 days. The importance of the extinction phase with cues cannot be overemphasized, as it tests whether self-administration can be supported by exposure to a cue that was previously associated with drug delivery. The phenomenon has been highlighted previously, especially in the C57BL/6 strain (Thomsen and Caine, 2011), and others have shown visual stimulus seeking in C57BL/6J mice, including the reinstatement of visual stimulusseeking even though no drug or food was ever paired with the stimulus (Contet et al., 2010). Drug reinstatement studies, which necessitate extinction without the discriminative cue, should therefore include the criterion of increased responding when cocaine is again made available after extinction (Thomsen and Caine, 2011). Interestingly, many studies have also capitalized on this interesting observation of operant sensation seeking in mice, using it as an additional tool to understand behavioral addictions more broadly (Olsen and Winder, 2009;Dickson and Mittleman, 2020).
Returning to our representative results, our dose-response findings indicate a peak in wildtype mouse responses at the 0.32unit dose, for which just over 75 responses were averaged in the 3-h session. This finding resembles wildtype mice published previously. For example, prior work shows wildtype C57BL/6 mice made ∼69 responses over a 3-h session (Thomsen et al., 2009a), and another showed ∼46 responses made over a 1.5h session (van der Veen et al., 2008); however, the peak of the average response is often alternatively seen at the 0.1 mg/kg/infusion dose, such as in a study reporting ∼60 responses in a 3-h session (Schmidt et al., 2011). Not many studies appear to have tested the mice under increased cost conditions comparable to the format and drug used in our example work, but our findings are very similar to our prior published work, where wildtype C57BL/6N mice earned around 19-24 infusions during 3-h sessions testing FR1, FR2, and FR3 schedules at this same cocaine dose (Penrod et al., 2020). These comparisons highlight that our results are within the normal limits found in the literature-or deviate in an expected way, such as due to more stringent extinction criteria-but we stress that a range of results can be observed across the literature for all self-administration phases or tests and may relate to differences in the mouse strain, timeout periods and maximum allowed responses, position and salience of discriminative cues, presence of unintended cues (such as syringe pump sounds), differences in chamber cleaning practices, etc.
The drug self-administration assay has high face validity and is a valuable tool for studying the reinforcing effects of drugs of abuse and substance use disorders. It can be adapted to investigate a wide range of drugs and drugrelated behaviors, including initial drug-taking, changes in sensitivity or tolerance, and persistence or reinstatement of drugseeking. While drug IVSA in mice presents challenges, there are tremendous advantages to being able to use the mouse research model in this assay, including the availability of genetic options, reduced need for the colony and behavioral testing space, lower cage and supply costs, and the ability to maintain continuity with prior or associated studies performed in mice.
Here we present an update to an established methodology for mouse IVSA supported by data in wildtype C57BL/6N mice that compares favorably to other published studies. Our hope is that this detailed protocol for catheter construction, surgery, equipment setup, and basic experimental planning may allow for more researchers to bring the advantages of the mouse model to bear on the problem of understanding the substance abuse.

DATA AVAILABILITY STATEMENT
The original contributions presented in the study are included in the article/supplementary materials, further inquiries can be directed to the corresponding author.

ETHICS STATEMENT
The animal study was reviewed and approved by Texas A&M University Institutional Animal Care and Use Committee.

AUTHOR CONTRIBUTIONS
The article was written by GV, JH, EC, CK, and LS. Contributions to the methodology were made by YG, JH, and LS. Surgery photographs were taken by GV. Behavioral data were collected by JH. All authors contributed to the article and approved the submitted version.

FUNDING
This work was supported by NIDA (DA051727) and Texas A&M University internal funding.