Summary Basis of Decision for Nplate ™

3.2.1 Pharmacodynamics

Pharmacodynamic (PD) studies were conducted with romiplostim utilizing doses that permitted the evaluation of a full range of dose-response activity. As there are currently no available animal models which are entirely analogous to ITP in humans, the clinical dose could not be selected directly from the animal pharmacology data.

Romiplostim has been evaluated in vitro and/or in vivo in several animal models including mice, rats, rabbits, and non-human primates (rhesus monkeys, cynomolgus monkeys, and baboons).

Romiplostim is a thrombopoiesis stimulating peptibody that increases platelet production through the activation of the TPO receptor (c-Mpl). Romiplostim mimics the activity of eTPO by binding to c-Mpl and transducing a similar cascade of intracellular events, thereby triggering the same biological outcome.

In primary PD studies, the binding ability of romiplostim was evaluated using Biacore, a technology used for studying biomolecular interactions. Results demonstrated that the dissociation constant (K_D) of TPO for human-c-Mpl was 33 nM, whereas the K_D of romiplostim for human-c-Mpl fell in the range of 14 to 51 nM.

In vitro studies demonstrated that romiplostim effectively stimulates megakaryocyte colony-forming cell proliferation from cynomolgus monkey, baboon, and human haematopoietic progenitor cells. Secondly, romiplostim binds to rat, cynomolgus monkey, and human platelets. Binding to human and monkey platelets was adequate and fully displaceable. For monkeys and humans, respectively, there was an 11-fold and 17-fold lower binding affinity when compared TPO. Furthermore, a study evaluating human platelet function in vitro found that at concentrations of 10 or 100 ng/mL romiplostim increased the adenosine diphosphate-generated aggregation response of platelets in a similar manner to the positive control megakaryocyte growth and development factor (MGDF), a c-Mpl ligand. The data suggest that the effects of romiplostim to stimulate platelet production are mediated via c-Mpl.

Romiplostim was screened in vitro for receptor binding affinity to sixty-three different receptors, enzymes, or ion channel targets. Results indicated that it is unlikely that romiplostim would cause any substantive pharmacologic activity in vivo other than the desired pharmacodynamics.

When tested in vivo, the administration of romiplostim caused increases in platelet counts in mice, rats, and non-human primates. These increases were dose-dependent, but dosing-route independent. The response in non-human primates was less pronounced than that seen in rodents, requiring higher doses to elicit a similar response.

Treatment with romiplostim in a mouse model of idiopathic ITP was effective in pre-splenectomy animals, following splenectomy (in animals that responded to splenectomy), in non-responders, and in responding animals who later relapsed, hence supporting its development for the use of romiplostim to treat patients with ITP.

3.2.2 Pharmacokinetics

Enzyme-linked immunosorbent assay (ELISA) assays were developed and validated for measuring romiplostim in the serum of rats, rabbits, rhesus monkeys, and cynomolgus monkeys. All relevant assays were shown to be sensitive, specific, reproducible and accurate enough for the assessment of pharmacokinetic (PK) and PD studies, and to support the toxicology studies of romiplostim in the relevant species.

Romiplostim was administered as a single intravenous (IV) administration to normal mice [30, 100 and 300 µg/kg bodyweight (bw)], neonatal Fc receptor (Fc-Rn)-deficient mice (100 and 1000 µg/kg bw) and thrombocytopaenic mice (3 and 30 µg/kg bw). In normal mice, the PKs of romiplostim were non-linear. Clearance and the volume of distribution at steady-state decreased with increasing dose with serum half-life values ranging from 6.1 hours (low dose) to 15.0 hours (high dose). Platelet response was dose-dependent and maximum platelet counts correlated with overall exposure to romiplostim. In Fc-Rn deficient mice, romiplostim was eliminated much faster than in normal mice, indicating that the platelet response was more pronounced in normal mice. This data suggest that the Fc-Rn receptor acts as a salvage receptor for romiplostim by maintaining romiplostim serum levels. In thrombocytopaenic mice, the effect was dose-dependent.  For example, at 3 μg/kg bw, thrombocytopaenic mice had a higher exposure than normal mice, while at 30 µg/kg bw, similar exposures were observed in normal and thrombocytopaenic mice suggesting that platelets play a role in the elimination of romiplostim, through a saturable process of target-mediated elimination.

Romiplostim was also administered as a single IV administration to normal rats, splenectomised rats, or bilaterally nephrectomised rats, at dose levels of 30, 100, 300 and/or 1000 µg/kg bw. The PK and PD profiles of romiplostim in control, sham-operated and splenectomised rats were similar, indicating that the spleen is not involved in the removal of romiplostim. In nephrectomised rats, exposure to romiplostim was significantly higher when compared to normal or sham-operated animals. In addition, the estimated apparent half-lives were significantly longer in nephrectomised animals than in sham-operated animals, suggesting that the kidney plays a significant role in the clearance of romiplostim during the first 12 hours.

Single dose IV and SC PK studies were conducted in rats, cynomolgus monkeys, and rhesus monkeys over a dose range of 30 to 300 µg/kg bw in rats and 500 to 5000 µg/kg bw in monkeys. After a single SC dose, bioavailability of romiplostim was approximately 21 to 28% in rats across the 30 to 300 µg/kg bw dose range and 19% in cynomolgus monkeys at 5000 µg/kg bw. For rhesus monkeys, bioavailability decreased from 74% at 500 µg/kg bw to 45% at 5000 µg/kg bw. Maximum concentration was achieved at a median time of peak concentration (t_max) of 8 to16 hours and 4 to 8 hours post-dose in rats and monkeys, respectively. The apparent terminal elimination t_½ was 19 to 21 hours in rats. In monkeys, the terminal elimination t_½ was 296 hours and 110 to 195 hours, for cynomolgus and rhesus monkeys, respectively, representing a slower absorption process into the systemic circulation. The single IV dose PKs were dose proportional for all three species. The estimated systemic clearance was approximately 6.3 to 8.4, 17.3 to 20.1 and 7.4 to 7.8 mL/hr/kg bw in rats, cynomolgus monkeys, and rhesus monkeys, respectively. The terminal elimination t_½ was approximately 17 to 19 hours in rats, 68 to 96 hours in cynomolgus monkeys, and 102 to 143 hours in rhesus monkeys.

Tissue distribution studies revealed that romiplostim distributed into all tissues. The highest concentrations of radioactivity were observed in the thyroid, blood components, kidneys, bone marrow, and liver at 0.5 hours post-dose. The highest percent values were seen in the blood, liver, muscle, skin and bone. Concentrations of radioactivity declined slowly and were detectable in all tissues at all collection time points (up to 168 hours). Radioactivity was detected in the brain at low levels, suggesting that the drug-derived radioactivity crossed the blood/brain barrier. Most of the radioactivity associated with serum, spleen, kidneys, heart, ovaries and liver was trichloacetic acid (TCA)-precipitable at all time points. Serum contained over 91% TCA-precipitable radioactivity throughout the study, and analyzed tissues exceeded 83% at all time points. It was determined that urine was the primary route of elimination with 87.7% of the administered dose being excreted in the urine and 6.56% excreted in faeces. However, only 10.7 to 11.3% of the radioactivity in urine and 45.9 to 49.2% of the radioactivity in faeces was TCA-precipitable.

Multiple dose SC PKs were evaluated in rats, cynomolgus monkeys, and rhesus monkeys. The test material was administered over a dose range of 10 to 100 µg/kg bw in rats, 500 to 5000 µg/kg bw in cynomolgus monkeys, and 100 to 5000 µg/kg bw in rhesus monkeys.

For rats, there was no appreciable sex difference in concentration profiles, and the sex-combined data showed no apparent accumulation over a 4-week dosing period. Exposure increased about 8- to 11-fold over the range of 30-100 µg/kg bw. Most of the serum samples in the 10 µg/kg bw group fell below the assay quantification limit. The incidence of anti-romiplostim binding and neutralizing antibodies was substantial, however, anti-TPO antibodies were not observed. When pregnant rats were dosed with romiplostim from gestational days 7 to 19 over the dose range of 10 to 100 µg/kg bw, there was no accumulation of romiplostim. The test material readily transferred across the placenta to the foetus, possibly mediated by the Fc-Rn-receptors. The average foetal serum concentration was approximately one-half of the average maternal serum concentration. A greater than dose proportional increase in area under the concentration-time curve (AUC) and the observed maximum serum concentration (C_max) was observed in the maternal and foetal serum. The romiplostim concentrations in the amniotic fluid were more variable relative to serum. The exposure in amniotic fluid only represented a fraction of the maternal serum exposure, that is to say (i.e.), 44% at 10 µg/kg bw, 17% at 30 µg/kg bw and 11% at 100 µg/kg bw. Exposure in pregnant rats appeared to be lower than in non-pregnant rats.

For monkeys, there were no appreciable sex differences in exposure. Exposure over a one-month dosing period increased less than dose proportionally for rhesus monkeys and increased approximately dose proportionally in cynomolgus monkeys. Exposure in rhesus monkeys was approximately twice as high as that observed in cynomolgus monkeys. The estimated accumulation ratios (ARs) were similar in rhesus monkeys and cynomolgus monkeys which ranged between 1.40 and 2.11 for the dose range of 300-5000 µg/kg bw. At 100 µg/kg bw, the exposure decreased following one month of dosing (AR = 0.437) which was most likely related to the elevated platelet counts. This hypothesis is supported by the finding that platelets play a role in the elimination of romiplostim and this role decreased with increasing dose. Neutralizing antibodies to romiplostim were not observed although binding antibodies occurred in both cynomolgus and rhesus monkeys. The presence of binding antibodies had no influence on the kinetics in cynomolgus monkeys, whereas the influence of binding antibodies could not be assessed in rhesus monkeys as all rhesus monkeys developed antibodies only after the treatment phase was completed.

Exposure of cynomolgus monkeys over a 3- or 6-month period was approximately dose proportional with no appreciable sex differences. The AR over the dosing period was approximately 1.05-1.65. The observed maximum serum concentrations occurred at 4 hours. The incidence of both binding and neutralizing antibodies to romiplostim were observed. Binding antibodies did not have an apparent influence on exposure. One monkey was found to have neutralizing antibodies on week 4 and had a reduced exposure during Week 13. No anti-TPO antibodies were observed.

3.2.3 Toxicology

All definitive toxicology studies utilized the SC route of administration as this is the intended clinical route of administration. Doses chosen were those that would provide a substantial challenge to the test animals and elicit systemic effects.

Single-Dose Toxicity

The toxicity of romiplostim was evaluated when administered as a single-dose via SC injection in Sprague-Dawley rats. The acute non-lethal dose of romiplostim was determined to be >1000 µg/kg bw (the highest dose tested), which represents a dose 33-fold greater than the highest anticipated clinical dose. All animals survived the duration of the study period. Treatment-related findings were considered non-adverse and reflected the pharmacological activity of the test material, i.e., increased platelet count, increased spleen weight, and increased splenic extramedullary haematopoiesis.

Cardiovascular

A single-dose SC cardiovascular safety study in cynomolgus monkeys at dose levels of 500, 1000 or 5000 µg/kg bw did not reveal any treatment-related effects on blood pressure, heart rate or body temperature. All electrocardiograms (ECGs) were qualitatively and quantitatively within normal limits. There were no alterations in morphology, rhythm or ECG intervals attributable to romiplostim.

Central Nervous System

A single-dose SC central nervous system (CNS) safety study in rats at dose levels of 10, 30 or 100 µg/kg bw did not reveal any treatment-related effects on the functional observational battery, body temperature, or motor activity.

Repeat-Dose Toxicity

Sub-chronic testing was conducted in rats, rhesus monkeys, and cynomolgus monkeys. Doses chosen provided continuous serum concentrations that were sufficiently higher than human concentrations, thus allowing for a meaningful evaluation of safety.

Sub-chronic testing in Sprague-Dawley rats was carried out for 1 month at dose levels of 0, 10, 30 or 100 µg/kg bw. Treatment-related mortality was evident at all dose levels tested. Although the specific cause of death could not definitively be determined, it was considered to most likely be the result of megakaryocytic thrombosis and increased blood viscosity related to the massive increases in platelet counts. Additional treatment-related findings (observed at all dose levels) were primarily related to the pharmacological activity of romiplostim and included: increased platelet count, increased spleen weights with correlating histopathological findings in the spleen (megakaryocyte hyperplasia, lymphocytic depletion), liver (megakaryocytosis, extramedullary haematopoiesis), lungs (megakaryocytosis) and bone marrow (megakaryocyte hyperplasia, myelofibrosis and hyperostosis). It was also noted that there was a generalized stimulatory effect on leukocyte production and a slight decrease in red blood cell count, haematocrit and haemoglobin. However, the values for the affected parameters still fell within the normal range. Platelet aggregation was slightly reduced for females at all dose levels, and for males at dose levels of 30 µg/kg bw and higher. In addition, there was a dose-related increase in the incidence of dermal, muscle, and subcutis chronic inflammation at the injection site. This was considered to be the result of slight trauma associated with the injection procedure and/or minor local irritation. All treatment-related findings were reversed following a 1-month recovery period. A high incidence of anti-romiplostim neutralizing antibodies was observed at doses ≥30 µg/kg bw. Consequently, it was deemed more appropriate to conduct longer-term studies in non-human primates as neutralizing antibodies rarely develop in these species. It is important to note that thrombocytopaenia was not observed in antibody-positive (Ab+) rats, indicating that antibody to romiplostim did not cross-react with eTPO.

Myelofibrosis was also observed in the 1-month rat study and was dose dependent both in terms of incidence and severity. However, the fibrosis resolved after cessation of drug administration and there were no clinical sequelae of this finding for the period observed.

Sub-chronic testing in cynomolgus and rhesus monkeys was carried out for 1 month as well as for 3 and 6 months in cynomolgus monkeys. The effective doses required in non-human primates to elicit a similar response to romiplostim as that seen in humans and rats were substantially higher. Hence, the dose levels chosen for these studies were 500, 1000 and 5000 µg/kg bw. Dose levels were generally well-tolerated and findings were consistent after 1, 3, or 6 months of study. Treatment-related findings were directly or indirectly related to the thrombopoietic activity of romiplostim. Stimulation of platelet production in the bone marrow was evident from microscopic observations of megakaryocyte hyperplasia and from the marked thrombocytosis that occurred at all dose levels. There was a dose-dependent decrease in red blood cell counts and related parameters, but the degree of anaemia was not clinically significant at any dose level. Increased spleen weight was noted at all dose levels and was of uncertain toxicological significance since there were no corresponding histopathological findings. Myelofibrosis was not observed in monkeys.

In addition, at the SC injection sites there was a dose-related increase in the incidence and severity of perivascular mononuclear cellular infiltrate. In general, all of the test article-related changes were completely or partially reversed during the 4-week (1-month study) or 8-week (3- and 6- month study) recovery period.

Genotoxicity

Romiplostim is composed entirely of naturally-occurring amino acids and contains no organic linkers or other non-protein portions. This suggests that it is unlikely that romiplostim would react directly with deoxyribonucleic acid (DNA) or other chromosomal material. Therefore, consistent with current ICH guidelines for genotoxicity testing for biotechnology-derived pharmaceuticals, these studies were not conducted.

Carcinogenicity

Studies to assess carcinogenicity of romiplostim were not conducted. However, as mentioned above, it is unlikely that romiplostim would react directly with DNA or other chromosomal material, therefore, mutagenic potential would not be anticipated.

Since romiplostim is highly immunogenic in rats, a traditional chronic oncogenicity study could not be conducted. However, it was considered unlikely that romiplostim would induce tumour development or growth since its mechanism of action is stimulation of platelet production via the Mpl receptor. In addition, there was no evidence of hyperplasia or pre-neoplastic lesions in the 6-month monkey study.

It was noted that there is a theoretical concern relative to the tumour stimulation capacity of TPO, and therefore of romiplostim, for certain subsets of acute myeloid leukemia (AML) cells.

Although it was stated that stimulatory effects were seen at TPO concentrations ranging from 20 to 200 ng/mL, which exceed the concentrations expected in the clinical studies, cell proliferation studies were not conducted by the sponsor. In the absence of an appropriate validated in vivo animal model for ITP, or an animal homologue for romiplostim, there remains concern regarding the long-term carcinogenic potential of romiplostim. For this reason, patients should be carefully monitored, and follow-up long-term safety data should be submitted when available.

Reproductive and Developmental Toxicity

A study was conducted to determine the effects of romiplostim on fertility when administered by SC injection to Sprague Dawley rats. Doses of 0, 10, 30, and 100 µg/kg bw were administered three times per week. Mortality was observed at all dose levels. All deaths were considered to be due to a combination of the pharmacological effect of romiplostim (increased platelet count) and the bleeding procedure. When the blood collection procedures were amended, no other deaths associated with blood collection occurred. Treatment with romiplostim did not have any effect on reproductive performance or fertility at any dose level tested.

Embryo-foetal developmental studies were conducted in mice and rats at dose levels of 0, 10, 30, and 100 µg/kg bw. For mice, romiplostim was embryo/foetal toxic in the absence of maternal toxicity at the highest dose of 100 µg/kg bw, based on an increased incidence of post-implantation loss. For rats, there were no adverse maternal or foetal effects noted at any dose level tested.

A pre- and post-natal developmental study was conducted in Sprague Dawley rats at dose levels of 0, 10, 30, and 100 µg/kg bw. Treatment of the initial parent generation (F₀) dams with romiplostim induced neutralizing antibodies to the test material in approximately 70% to 80% of the dams in each group. This created two subgroups, Ab+ and antibody-negative (Ab-), which were evaluated separately and compared as appropriate. Treatment-related maternal mortality was noted at 30 and 100 µg/kg bw shortly after blood collection procedures, which was considered secondary to the increased blood viscosity resulting from markedly elevated platelet counts. As well, an increase in the incidence of stillborn pups was noted in the 100 µg/kg bw group for both the Ab+ and Ab- subgroups. There was also an increase in pup loss during the early lactation period. A total litter loss also occurred at 30 µg/kg bw, in the Ab+ subgroup, contributing to the incidence of both stillborn and early post-parturient decedent pups. Despite the small increase in peri-natal pup mortality, there was no notable effect on overall live litter size. Also, there was no effect on post-weaning functional development (open field assessment, motor activity, learning test), preputial separation/vaginal opening, estrous cycling or gross necropsy.

Based on the data from the reproductive and developmental studies conducted, it is recommended that special consideration be given to women who are being treated with romiplostim who may wish to become pregnant. It is recommended that romiplostim not be used during pregnancy unless the potential benefit to the mother justifies the potential risk to the foetus or embryo.

Local Tolerance

Local tolerance studies were not conducted; however, no evidence of substantive irritation was identified at the sites of injection in the repeat-dose toxicity studies. As such, it is unlikely that the formulated romiplostim would pose a meaningful risk for reaction at the site of injection.

Antigenicity

In an antigenicity study conducted in mice, it was determined that although 60 to 80% of the mice generated an antibody response to romiplostim within 2 weeks of a single exposure, endogenous platelet production was not impacted within the period evaluated. The strongest serum interactions were to thrombopoietin mimetic peptide (TMP).

Subsequent exposure to romiplostim resulted in reduced efficacy to lower doses of romiplostim. Increased doses of romiplostim were able to stimulate a platelet response even in the presence of antibodies to romiplostim.

Overall, it appears that anti-romiplostim antibodies may reduce the effectiveness of the drug, but do not lead to thrombocytopaenia, which would be the expected outcome if these antibodies cross-neutralized eTPO.

3.2.4 Summary and Conclusion

The non-clinical PD and PK data provided support for the mechanism of action of romiplostim to increase platelet production through the activation of the TPO receptor c-Mpl.

Non-clinical toxicity testing demonstrated that romiplostim was generally well-tolerated in rats, mice, cynomolgus monkeys and rhesus monkeys. Treatment-related findings were primarily related to the pharmacological activity of the test material. The non-clinical toxicology data base was considered adequate to assess the safety profile of romiplostim and to support its use in humans, provided adequate safety precautions are taken, as described above.