This is the first human clinical study using peg-rhArg1 in the treatment of advanced human malignancy. We have demonstrated that arginine depletion in humans can be achieved safely with peg-rhAgr1 in a dose-response manner. There was no treatment-related death and all the AEs, except for the one possible case of DLT in cohort 4, were mild grade 1 and 2 non-hematologic toxicities, such as nausea and abdominal discomfort, which resolved spontaneously with symptomatic treatment. Thus, the primary objectives of the study were met and the recommended weekly dose for future development of Peg-rhAgr1 was chosen as 1600U/kg.
As in any drug trials involving patients with severely compromised liver functions, there is always a problem in attributing the exact cause of dysfunction, whether due to the test drug or progression of underlying liver disease, i.e. cirrhosis, hepatitis or malignancy, often all three. Overall, only 2 of the 15 patients in the entire study had possible severe drug-related liver dysfunction, one in cohort 1 and the other in cohort 4. The first patient had grade 3 elevation of ALT/AST after the first injection and was taken off the study. Before he went into hepatic encephalopathy with liver failure, a CT scan confirmed progression of the liver tumour, from which he died after 2 months off treatment. The second case had grade 3 followed by 4 hyperbilirubinemia after a single injection of peg-rhArg1. Since the patient had no overt clinical signs of disease progression, this hyperbilibrubinemia was deemed as DLT. Although great care was taken to recruit patients with adequate life-expectancy into the study, the poor outcomes of these patients early on in their treatment illustrate how unpredictable the clinical course of heavily pre-treated HCC patients with advanced disease can be. In particular, nearly half of the recruited patients were heavily pre-treated with various prior system therapies, which might account for the rapid deterioration of the recruited patients in the study. As per protocol requirement, the additional 3 cases were recruited into cohort 4, and it was reassuring that no further liver DLT was observed. The study was terminated on completion of all the 6 cases in cohort 4 since the MTD and OBD—the two important objective parameters the authors aimed to establish—had already been found.
In terms of toxicity profiles, peg-rhAgr1 does have an advantage over both ADI and sorafenib. When compared with ADI, peg-rhAgr1 releases urea as a side product, which is much less toxic than the ammonia that is released by ADI treatment. This is important since in patients with compromised liver function, as in most HCC cases, the extra ammonia may make the patients more susceptible to develop HE. Indeed, the most common AEs attributed to ADI include transient and reversible HE. Other toxicities reported in ADI included skin and injection site necrosis and electrolyte disturbance, which were not seen in our study. Being a Mycoplasma product, ADI is a xeonbiotic immunogenic substance, despite pegylation. It has been reported that neutralizing antibodies to ADI developed ~50 days after initiation of treatment . Peg-rhAgr1, on the other hand, is a recombinant human hepatic enzyme and in our preclinical and clinical studies (data not shown); no evidence of immunogenicity has been found to date. This has important therapeutic implication since the neutralizing antibodies would render the drug ineffective as an arginine-depleting agent. It was also reported in the ADI studies that arginine levels would rise in the second month of treatment, coinciding with the development of neutralizing antibodies to ADI. On the other hand, neutralizing antibodies to Peg-rhAgr1 have not been detected thus far in the present study and it is reassuring that arginine levels remained at lower levels throughout the entire study as long as treatment continued. More importantly, the AEs profile of peg-rhAgr1 also compares favourably with that of sorafenib, which has large numbers of grade III and IV toxicities in terms of hand-foot-skin reaction, hypertension and rash. Again, none of these were seen in this study.
Enzymatic depletion or deprivation therapy, whether with peg-rhAgr1 or ADI, has now emerged as a new platform of anti-cancer treatment in a number of human malignancies, such as HCC and melanoma . In vitro and in vivo xenograft studies suggest that this mode of treatment is efficacious also in prostatic and renal cell cancers, in which ASS and arginosuccinate lyase (ASL) were highly repressed or absent in an MD Anderson series . Pegylated human arginase has recently been shown in one US study to induce remission in acute T cell leukemia . The present study serves to prove the concept that arginine depletion is really a promising approach in treating advanced HCC patients. Thus, further evaluation of peg-rhArg1 in future phase-II trials of advanced HCC is warranted.
Peg-rhAgr1 might have demonstrated early activity in the group of heavily pre-treated advanced HCC patients. Of the 8 cases completed 8 weeks of treatment and beyond, 4 had durable SD lasting for 3 m+, 6 m+, 6 m+ and 3 m+, respectively with drop in AFP level. One patient had minor shrinkage of the target lung lesions that did not qualify as a partial response by the RECIST 1.0 criteria. As experienced by other clinical investigators in the field of HCC, assessing CT response to treatment in HCC is notoriously difficult and tumour response may result in tumour necrosis instead of shrinkage (Fig. 2) . Notably, at the time when this study was conducted, RECIST 1.0 criteria was the standard for assessing tumor response . As yet, there is no consensus in the HCC community about which new radiological criteria, such as EASL  or modified RECIST  would be a better method to assess response in HCC patients receiving targeted therapy or biological therapy alone. Unfortunately, 7 enrolled patients were not evaluable for treatment response and most of these patients discontinued the therapy early mainly due to either rapid disease progression or various complications from underlying cirrhosis. This indeed reflects the challenges in conducting clinical trials in this refractory patient population with aggressive tumour biology and poor liver reserve.
In conclusion, peg-rhArg1 has a manageable safety profile and is potentially a superior arginine depleting agent than ADI in the treatment of human malignancies due to its low toxicity profile and sustainable arginine depletion. The optimal therapeutic dose of peg-rhArgI is 1600U/kg in humans for future studies in advanced HCC patients.