EXAMPLES
Example 1
Acetylsalicylic Acid and Acetaminophen

Preparation of Formulation Stocks:

Acetylsalicylic acid stock (30 mM) is prepared by adding 74.9 mg of solid to 13.9 mL of water. The acetylsalicylic acid stock solution is solubilized by the addition of a small aliquot (<50 uL) of 5 M NaOH to increase the pH to 12.5. Paracetamol stock (90 mM) is prepared by adding 101.7 mg of solid paracetamol to 7.5 mL of water (pH=5.3). The phosphate buffer stock (700 mM) is prepared by adding 1.5 g of Na₂HPO₄-7H₂O to 8.1 mL of water (pH 9.0). The m-cresol stock solution (226.6 mM) is prepared by adding 49 mg of m-cresol to 2 mL of PBS. The zinc stock (30 mM) is prepared by adding 50.2 mg of zinc oxide (ZnO) to 20.6 mL of 75 mM HCl.

Insulin Co-Formulation Preparation:

Insulin lispro control formulation (750 uL) is prepared with a final concentration of 0.6 mM insulin lispro (3.5 mg/mL; 137.7 uL of 3267 uM Zn²⁺free stock), 29.13 mM m-cresol (3.15 mg/mL; 96.4 uL stock), 7 mM Na₂HPO₄-7H₂O (7.5 uL of stock, pH 9.0), and 0.3 mM ZnO (7.5 uL of stock; 3 mol Zn²⁺/hexamer). The insulin lispro/acetylsalicylic acid formulations are prepared as the control formulation above with a supplementation of 10 mM or 20 mM acetylsalicylic acid as final concentrations (250 uL and 500 uL of stock, respectively). Additionally, insulin lispro/paracetamol formulations are prepared as the control formulation above with a supplementation of 30 mM or 60 mM paracetamol as final concentrations (250 uL and 500 uL of stock, respectively). Two separate formulations are prepared as described above with 20 mM acetylsalicylic acid or 90 mM paracetamol; however, the m-cresol is excluded from the formulation. All samples are adjusted with water to achieve the final concentrations. The pH is adjusted to 7.4+/−0.1 with dilute HCl and/or NaOH. All samples are filtered using 0.22 um syringe filter (Millex-GV; REF SLGV013SL). An aliquot of 300 uL of each formulation described above, as well as an aliquot of 300 uL of commercial Humalog®, are placed into separate 2-mL screw-cap glass vials and incubated static at 37° C. The insulin lispro main peak loss and high molecular weight polymer (HMWP) growth over time are tracked using analytical reversed-phase HPLC for 24 days. The samples are visibly inspected for particulate or precipitation throughout study.

Analytical RP-HPLC Conditions:

Insulin lispro main peak loss and HMWP growth are tracked over time using HPLC. Mobile phase buffer A (100 mM sodium sulfate+10% acetonitrile, pH to 2.3) and mobile phase buffer B (80% acetonitrile in water) are filtered through 0.22 um filter. The gradient is presented in Table 1. The UV detector is set to 214 nm wavelength, column temperature to 40° C., flowrate to 1 mL/min, and injection volume to 5 uL of each formulation. Main peak and HMWP are reported as an area percent from main peak region at approximately 10 mins retention time and HMWP region between 17 and 29 mins. The data is collected and slope analysis of rate of change per day is performed using Y by X plots using commercial software.

TABLE 1

Analytical RP-HPLC assay for determination of

insulin main peak loss and polymer formation.

Time (min)
Buffer A (%)
Buffer B (%)

0.00
76.8
23.2

3.00
74.5
25.5

15.00
74.5
25.5

21.00
72.4
27.6

26.00
51.8
48.2

27.00
51.8
48.2

27.10
76.8
23.2

35.00
76.8
23.2

TABLE 2

Tabulated rates of change (%/day) for main peak loss and HMWP

growth over time as determined by analytical RP-HPLC.

Change in Main Peak
Change in HMWP

Sample
(%/Day)
(%/Day)

10 mM Aspirin
−0.12 +/− 0.05
0.03 +/− 0.01

20 mM Aspirin
−0.17 +/− 0.11
0.09 +/− 0.07

20 mM Acetylsalicylic
−1.47 +/− 0.26
0.84 +/− 0.17

acid/No m-cresol

30 mM Paracetamol
−0.33 +/− 0.22
0.14 +/− 0.03

60 mM Paracetamol
−0.37 +/− 0.01
0.27 +/− 0.05

60 mM Paracetamol/No
−0.94 +/− 0.03
0.50 +/− 0.02

m-cresol

Insulin lispro control
−0.16 +/− 0.05
0.04 +/− 0.003

Commercial Humalog
−0.18 +/− 0.04
0.04 +/− 0.01

The tabulated data (Table 2) summarizes the rates of change (0%/day) in main peak and HMWP growth for insulin lispro co-formulated with either acetylsalicylic acid or paracetamol, in the absence and presence of m-cresol. For comparison, commercial Humalog® and laboratory prepared insulin lispro control, based on the Humalog® formulation, are presented as controls. The inclusion of acetylsalicylic acid or paracetamol has minimal impact on insulin lispro stability. The absence of m-cresol results in a higher rate of insulin main peak loss and polymer formation compared to m-cresol containing formulations.

Example 2
Heparin

Insulin with heparin formulations are prepared by adding 0.3 mL of heparin sodium (10,000 U/mL in water) and 0.58 mL of 0.9% sodium chloride to 10 mL of U110 insulin lispro to provide a final formulation comprising 276 U/mL heparin sodium, 101 U/mL insulin lispro, 16.22 mg/mL glycerin, 3.19 mg/mL m-cresol, 1.91 mg/mL sodium phosphate dibasic heptahydrate, 0.46 mg/mL sodium chloride, and 0.02 mg/mL Zn²⁺. The insulin/heparin solution is transferred to pump cartridges for use in the pre-clinical in vivo model.

Pre-Clinical In Vivo Model:

A preclinical model using non-diabetic, castrated male Yucatan mini pigs in non-terminal, randomized, cross-over fashion is useful for mimicking the behavior of insulin pharmacokinetics and glucose pharmacodynamics in human subjects over the course of continuous infusion at a single infusion site for up to 10 days. In this model, one infusion set is placed on the abdomen of each animal, and each site receives the same volume of insulin placebo (formulations identical to U100 Humalog® but containing no active pharmaceutical ingredient) for 3× daily bolus infusion of 0.1 mL, and continuous basal infusion of 0.01 mL/h. Test groups with non-insulin medicaments receive placebo with the medicament added at the desired concentration. On days 0, 3, 5, 7, and 10 following infusion set placement, after an overnight fast, each animal is given an infusion bolus of U100 Humalog (with or without a non-insulin medicament) in accordance with their body weight (0.15 U/kg). Glucose and insulin concentrations are then assessed via serial blood samples collected from each animal at the following time points relative to the time of insulin bolus delivery: −30, −20, 0, (deliver insulin bolus), 15, 30, 45, 60, 75, 90, 120, 150, 180, 240, 300, and 360 min. At conclusion of blood sampling on day 10, infusion sets are carefully removed and assessed for catheter damage. The site of infusion is also assessed for signed of infection, irritation, erythema, and edema. This model can demonstrate significant reductions in plasma insulin concentrations and consequent glucose lowering over time indicative of the loss of infusion site efficacy. The pre-clinical model is used to show the effects of compositions claimed herein.

The data in Table 3 represents the insulin pharmacokinetic data for U100 Humalog and the insulin/heparin co-formulation, respectively, and over the course of 10 days of continuous infusion at a single infusion site. The diminishing concentration of serum insulin over time for the Control Group of Table 3 is typical for CSII, while the maintenance of insulin pharmacokinetics without any loss in signal for 7 consecutive days for the insulin/heparin co-formulation Test Group in Table 3 supports the desired result.

TABLE 3

Maximum changes in plasma glucose concentration after dosing with

0.15 U/kg of U100 insulin lispro or 0.15 U/kg of U100 insulin lispro

spiked with 267 U/mL heparin sodium, on days 0, 3, 5, 7, and 10 of

continuous infusion at a single infusion site. Variation is standard

error of the mean. P-values derived via a 2-sample T-test.

Maximum Change in

Plasma Glucose

Days of

(Mean ± SEM)

Continuous
Control Group
Heparin-Treated
P-value

Infusion
(N = 6)
(N = 7)
(α = 0.05)

0
−71.8 ± 5.2
−66.9 ± 2.6
0.416

3
−57.0 ± 6.0
−65.1 ± 2.7
0.261

5
−42.2 ± 8.0
−62.0 ± 2.8
0.057

7
−36.2 ± 10.3
−62.9 ± 5.3
0.055

10
−39.1 ± 10.4
−22.3 ± 8.7
0.247

The effect on insulin efficacy is further demonstrated by evaluating the maximum serum insulin concentrations over time. This study showed that the glucose lowering effect of an insulin bolus is sustained for 7 days in heparin-treated animals, but steadily declines from day 0 in animals not treated with heparin. N=6 for each group.

Example 3
Dexamethasone

This study can be used to evaluate the beneficial effect of local delivery of dexamethasone during SCII. In this experiment, dexamethasone is not co-formulated with insulin; rather, the 0.25 mg/h infusion is achieved using a commercially available solution of 10 mg/mL dexamethasone delivered using an infusion pump programmed for a basal delivery of 25 uL/h. On days 0, 3, 5, 7, and 10 following infusion set placement, the animals are fasted followed by dosing with 0.15 U/kg insulin lispro using an infusion pump. During the time between insulin doses, the animals receive 25 uL/h continuous infusion of either 10 mg/mL dexamethasone or insulin placebo.

Results using the methods set forth support that local delivery of dexamethosone via co-formulation with insulin is a preferred route of administration to achieve the desired result. Results from the study support that the concentration of 0.25 mg/h (6 mg/d) is effective in improving infusion site viability for 5 days.

Dexamethasone insulin formulations are prepared substantially as follows: 1.1 mL of a 100 mg/mL dexamethasone solution in sodium phosphate buffer is added to 10 mL U110 insulin lispro. The formulation is prepared using standard techniques to provide a formulation comprising 10 mg/mL dexamethasone, 99 U/mL insulin lispro, 15.84 mg/mL glycerin, 3.12 mg/mL m-cresol, 2.07 mg/mL sodium phosphate, and 0.02 mg/mL Zn².

Results shown in Table 4 indicate that dexamethasone can improve infusion site viability for up to 5 days, whereas orally delivered dexamethasone has no effect.

TABLE 4

Baseline-adjusted areas under the curve for plasma glucose over the 6 hours following

an insulin bolus of 0.15 U/kg of U100 insulin lispro, on days 0, 3, 5, 7, and

10 of continuous infusion of 0.25 mg/h dexamethasone (Subcutaneous Dexamethasone),

25 uL/h insulin placebo (Control Group), and 25 uL/h insulin placebo +

twice daily oral administration of 3 mg dexamethasone (Oral Dexamethasone).

Variation is standard error of the mean. P-values derived via a 2-sample T-test

between the Subcutaneous Dexamethasone and Control Group data sets.

Plasma Glucose

Days of
Control
Subcutaneous
Oral
* P-value vs.

Continuous
Group
Dexamethasone
Dexamethasone
Control

Infusion
(N = 5)
(N = 5)*
(N = 5)
(α = 0.05)

0
−15110 ± 2482
−16700 ± 2027
−14218 ± 754
0.635

3
−7015 ± 2418
−12591 ± 873
−5265 ± 2130
0.082

5
−8266 ± 1422
−12297 ± 755
−8109 ± 987
0.046

7
−5517 ± 1340
−7149 ± 1555
−7750 ± 1491
0.453

10
−4440 ± 1242
−4893 ± 775
−4266 ± 1551
0.767

Example 4

Meloxicam and Bromfenac Sodium

Preparation of Formulation Stocks:

Humalog U200 stock solution is prepared by dissolving 7.55 g of insulin lispro powder in approximately 788 g of mixture containing 3.15 g of m-cresol, 16 g of glycerin in water, supplemented with 1.96 mL of 12.44 mg/mL of zinc oxide stock solution. After adding insulin lispro powder, the pH of the solution is reduced to pH 3.0+/−0.2 by adding suitable amount of 10% w/w HCl and stirring until complete dissolution is observed or up to a maximum of 2 hours. Following insulin lispro dissolution, the pH of the solution is increased to pH 7.6+/−0.1 by adding suitable amount of 10% w/w NaOH, allowing the mixture to re-dissolve the solids for 30 minutes and then adjusting the pH down to pH 7.3+/−0.1. Then, 90 mL of Tris base prepared at 55.55 mM and pH 7.2-pH 7.4 in water is added to this solution After this, the solution pH is adjusted down to pH 7.25+/−0.5 by adding suitable volume of 10% w/w HCl. The solution is supplemented with additional water until its weight equals 915 g. The final solution is mixed for another 5 minutes and then filtered through a 0.22 um PVDF filter into a clean container.

Meloxicam stock (3.0 mg/mL) is prepared by adding 179.8 mg of solid first to 20 mL of water, then increasing the pH to above pH 12.0 using 10% w/w NaOH, followed by addition of more water until 59.8 mL of total water volume has been added. This stock solution is mixed for at least 30 minutes or until complete dissolution is observed.

Humalog U200-Meloxicam Co-Formulation Preparation:

The Humalog U220 stock solution is taken in clean glass beakers with a stir bars. Suitable volume of the meloxicam (3 mg/mL) stock solution is added along with suitable volume of water to equal 0.01, 0.05, 0.18, or 0.27 mg/mL meloxicam and mixed well. The pH of the solution is adjusted at pH 7.25+/−0.05 by adding suitable volume of 10% w/w HCl. Each solution is filtered through 0.22 um PVDF filters and then filled into clean glass vials. The vials are kept at 30° C. for 36 days.

Humalog U200-Bromfenac Sodium Co-Formulation Preparation:

The Humalog U220 stock solution is diluted with water to equal to U200/mL insulin lispro concentration. Suitable quantity of Bromfenac Sodium powder is weighed out in separate clean glass beakers with a target volume of 200-220 mL of solution containing 0.075, 0.187 and 0.746 mg/mL bromfenac sodium respectively. Based on the weighed quantity, the appropriate volume of Humalog U200 prepared earlier is added and mixed for 5 minutes to fully dissolve the bromfenac sodium. The resulting solution filtered through 0.22 um PVDF filters and then filled into clean glass vials. The vials are kept at 30° C. for 36 days.

Analytical SEC Conditions for Humalog U200-Meloxicam Coformulation Samples

The mobile phase is prepared by mixing a 1.0 mg/mL stock solution of L-arginine with acetonitrile and glacial acetic acid in the ratio 65:20:15 respectively. The mobile phase is flowed at 0.5 mL/min through a size-exclusion HPLC column maintained at ambient temperature (WAT201549, 30 cm×7.8 mm, Waters). 50 uL of each coformulation sample is injected and monitored for a run time of 40 minutes at 276 nm using a UV-Visible detector. Main peak and high molecular weight polymer (HMWP) are reported as an area percent from main peak region at approximately 18 mins retention time and HMWP region between 10 and 17 mins. The data is collected at different points during the 36-day incubation period at 30° C. Using slope analysis, the rate of change of HMWP per day is performed using Y by X plots using commercial software.

Analytical SEC Conditions for Humalog U200—Bromfenac Sodium Coformulation Samples:

The mobile phase is prepared by mixing acetonitrile and water in 1:1 volume ratio and further mixing trifluoroacetic acid (TFA) at 0.1% volume ratio in the solution. The mobile phase is flowed at 1.2 mL/min through a size-exclusion HPLC column maintained at ambient temperature (233080-7830, 300 mm×7.8 mm, Sepax Technologies). 25 uL of each coformulation sample is injected and monitored at 276 nm using a UV-Visible detector for 15 minutes. Main peak and high molecular weight polymer (HMWP) are reported as an area percent from main peak region at approximately 5.6 mins retention time and HMWP region between 4 and 5.2 mins. The data is collected at different points during the 36-day incubation period at 30° C. Using slope analysis, the rate of change of HMWP per day is performed using Y by X plots using commercial software.

TABLE 3

Tabulated rates of change (%/day) for main peak loss and HMWP

growth over time as determined by analytical RP-HPLC.

Change in HMWP

Sample
(%/Day)

Humalog U200 + 0.01 mg/mL meloxicam
0.0091

Humalog U200 + 0.05 mg/mL meloxicam
0.0091

Humalog U200 + 0.18 mg/mL meloxicam
0.0082

Humalog U200 + 0.27 mg/mL meloxicam
0.0079

Humalog U200 control
0.0079

Humalog U200 + 0.075 mg/mL bromfenac sodium
0.0079

Humalog U200 + 0.187 mg/mL bromfenac sodium
0.0089

Humalog U200 + 0.746 mg/mL bromfenac sodium
0.00122

Data Analysis Methodology

There are four parameters that we most often use to evaluate the effects of an intervention on glucose pharmacodynamics: the baseline subtracted area under the glucose PD curve (AUC); C_min, or the maximum change from baseline that the insulin dose effects; T_min, or the time it takes to reach C_minafter the insulin dose is administered; and the “Early 50% T_min”, or the time it takes to first reach a glucose concentration that is 50% the magnitude of C_min. Table 4 below represents typical PD data for a non-diabetic swine that has been fasted and dosed with 0.15 U/kg of U100 insulin lispro, delivered via an infusion set that has been in place for up to 10 days. Between insulin doses, the infusion set is connected to an insulin pump programmed to deliver insulin placebo at a 10 uL/h basal rate and 100 ul boluses delivered three times daily. SEM=standard error of the mean.

TABLE 4

Typical PD data for a non-diabetic swine

Days of

Std

CSI
N
Mean
SEM
Dev

C_min
0
34
−61.32
1.74
10.15

(Δ mg/dL)
3
33
−57.31
1.88
10.79

5
30
−58.16
2.37
13

7
29
−45.12
3.64
19.63

10
26
−30.44
3.18
16.23

T_min
0
34
69.26
7.33
42.75

(min)
3
33
52.73
6.69
38.45

5
30
40
3.9
21.33

7
29
69.3
17
91.5

10
26
96.9
20
101.9

AUC
0
34
−13196
602
3511

(min*mg/dL)
3
33
−9702
529
3039

5
30
−8277
620
3394

7
29
−6115
566
3049

10
26
−4213
509
2595

As is demonstrated above by the shift in PD parameters from day 0 to day 3, starting at around day 2 of continuous subcutaneous insulin infusion (CSII), insulin uptake becomes faster, and depending on pump speed, some loss in AUC may also be observed. This phenomenon has been well characterized by researchers (sometimes referred to as the “Tamborlane effect”) and is well-tolerated by patients. Thus, a solution to enabling extended infusion site efficacy should strive to achieve some level of equivalence to the glycemic control patients experience on day 3 of CSII. In the interventions set forth herein, we have thus evaluated the resulting PD performance against the standard of care (a rapid-acting insulin such as Humalog®) after 3 days of CSII. Data from swine studies set forth in Tables 5, 6, and 7 show the efficacy of three formulations at days 0, 3, 5, 7, and 10 normalized against U100 Humalog values on day 3.

TABLE 5

Glucose Area Under the Curve v. Standard of Care Day 3