Good morning. My name is Jaume Pons. I am the Chief Executive Officer of ALX Oncology. After the usual disclaimers in slide number one and some pictures of us and our background in slide number two, let's move to slide number three. ALX Oncology is a clinical-stage immuno-oncology company that is focused on the CD47 pathway. ALX148, our phase II clinical candidate, is a CD47 blocker that is designed to be used in combination. This design has shown in the clinic a very good tolerability profile that has enabled higher dosing, and in the clinic, this higher dosing has enabled greater efficacy. At this moment, we have clinical proof of principle, both in hematological and solid tumors. Our initial focus is on solid tumors, myelodysplastic syndrome, and acute myeloid leukemia.
In addition to ALX148, we have a second compound that we call SIRPα TRAAC that will be in the Investigational New Drug by the end of 2022. In slide number four, CD47 as a target is highly expressed in tumor cells here on the left in red, it's also highly expressed in normal cells in black. Therefore, if we use CD47 as a tumor-associated antigen, we may have problems with therapeutic window. By destroying cancer cells could also be destroying normal cells. What makes CD47 interesting to me is its role as a checkpoint modulator. Many anti-cancer agents, for example, anti-cancer antibodies, bind cancer cells and provide what we call an "eat me" signal. In that context, cancer cells upregulate CD47 that interacts with SIRPα on macrophages and other myeloid cells and sends what we call a "don't eat me" signal. SIRPα is an inhibitory receptor on myeloid cell.
It's a checkpoint for myeloid cells. The interaction between CD47 and SIRPα limits the activity of many anti-cancer agents. One way to treat CD47 as a target is making CD47 as a tumor-associated antigen, therefore designing molecules that bind CD47 and at the same time provide the required positive signal, the "eat me" signal. These kind of molecules can have single agent activity because they block the "don't eat me" signal and at the same time, they provide the required "eat me" signal. The problem is that CD47 is also expressed in normal cells. For example, T cells, platelets, and neutrophils. These approaches have shown in the clinic cytopenias related to this approach. These kind of molecules, therefore, they cannot be dosed high enough, and in the combination setting, what is required is to completely block the CD47/SIRPα pathway.
They cannot go high enough and completely block it, and therefore cannot enable the full activity of the anti-cancer drug. What we decided to do is design a CD47 blocker that treated CD47 only as a checkpoint modulator. We made a molecule that binds very strongly to CD47 and blocks interaction with SIRPα but does not have an active fragment crystallizable. It is not able to interact with the Fc gamma receptors, and therefore it does not provide the required positive signal. Our approach will not have single agent activity. At the same time, because we don't provide the positive signal, despite our molecule will bind normal cells, it does not destroy normal cells. For example, we have proven that we do not have any dose-related cytopenias in the clinic.
In that situation, when we combine with an anti-cancer agent, we can dose very high, completely block CD47 and SIRPα, and fully enabling the activity of the companion drug. In slide number seven, you have a cartoon of ALX148. Our CD47 binding domain is what we call high-affinity SIRPα. That's the extracellular domain of SIRPα, the receptor for CD47, that has been affinity matured to bind CD47 with very high affinity, with picomolar affinity. It's fused to a completely inactivated Fc that provides the same half-life of an antibody, the same pharmacokinetics, but does not interact with Fc gamma receptors. Therefore, it does not activate macrophages directly. Molecular weight is half an antibody, so if you want to compare our dosing with antibody dosing, you have to multiply by two.
10 mg/ kg of ALX148 is equivalent to 20 mg/ kg of an antibody in terms of binding sites. The smaller molecular weight may enable better tumor penetration. The molecule is cross-reactive to human, monkey, and mouse, so we can run engineered models, the clinical models are more relevant of the clinical setting. We have a completely antibody standard manufacturing process and very good stability. In slide number eight, you can see that in the combination setting, ALX148 has higher activity than the competition molecules with active Fcs. For example, in this example, we have cancer cells that express Epidermal Growth Factor Receptor and human macrophages.
We provide cetuximab that targets cancer cells, so provides the required positive signal and is the orange dot on the left. On top of that, we have a dose response of ALX148, paclitaxel, and TTI-621 that we make in-house based on public information. You can see that ALX148 with the highest affinity and the lowest receptor function is the molecule that enhances more the activity of cetuximab in this combination setting. Our molecule cross-reacts to mouse, we have been able to test safety in mouse as well. In this experiment, we're comparing ALX148 with an inactive Fc with the same binding domain of CD47, so with active Fc, that's ALX377. You can see that a single dose of ALX377 reduces red blood cells, platelets, and white blood cells, while ALX148 does not.
This experiment proved to us that the cytopenias seen in the clinic by others are directly related to the active Fc and is not a target effect by binding CD47. We have proven in preclinical models that we can enhance anti-cancer antibodies in xenograft models, for example, cetuximab on the left. We have proven as well that we can enhance checkpoint modulators in syngeneic models. In the right, for example, combination with atezolizumab and anti-PD-1 antibody, but we have also data with anti-PD-1 antibodies. Next slide, Elena. You can see with our pharmacokinetics profile that we have linear PK at doses of 3 mg/kg and higher. The combination of antibodies do not change PK, and the half-life at 3 mg/kg and higher is predicted about 30 days.
On the right, you see target occupancy on T cells, and you can see that we have full receptor occupancy across the dosing interval at 3 mg/kg weekly and higher. In terms of safety, we have been able to dose very high in preclinical models up to 100 mg/kg in non-human primates. That's equivalent to 200 mg/kg of antibody with no observable adverse events. As a single agent in humans, up to 30 mg/kg every other week, with no evidence of dose-dependent cytopenias. In combination, we have defined the phase II recommended dose of 15 mg/kg weekly, but in terms of exposure, because our safety profile is good and we can dose higher, we have never reached a maximum tolerated dose, we can increase the dosing to extend the dosing interval.
For example, if we combine with something that is dosed weekly, we can do 15 mg/kg weekly, or we can do 30 mg/kg every other week or 45 every three weeks if we're combining with pembrolizumab, or 60 every four weeks if we're combining with decitabine that is dosed monthly. Our safety profile allows a very flexible dosing schedule, maintaining the same exposure. In slide 13, you see a high-level description of our clinical progress so far. We have tested ALX148 in head and neck squamous cell carcinoma in combination with KEYTRUDA and combination with KEYTRUDA plus carboplatin. We have Fast Track designation for this indication. Now we're moving into phase II in first-line head and neck squamous cell carcinoma in combination with KEYTRUDA, and also another phase II in combination with KEYTRUDA plus chemotherapy.
In gastric cancer, we have tested ALX148 in the second line for patients that had failed HERCEPTIN in the first line, in combination with HERCEPTIN, in combination with HERCEPTIN plus chemotherapy. We also have Fast Track designation for this indication. We're also moving into phase II randomized, potentially for an accelerated approval. In collaboration with Zymeworks, we're testing ALX148 with zanidatamab, a bispecific anti-Human Epidermal growth factor Receptor 2 antibody in breast cancer. The phase I should start shortly. In hematologic malignancies, our focus is MDS and AML. In MDS, we are now in the middle of phase I in combination with decitabine to start a phase II this year. We're going to start a phase I in AML also this year. We have a very good proof of principle study in Non-Hodgkin lymphoma in combination with rituximab, patients that have had rituximab.
This is a study that allows us to compare ourselves with the data from others' CD47 blockers. Finally, we have a second program that I will describe later. In terms of safety in humans, it has been very solid. The most common side effect is fatigue and some rash. You can see that cytopenias like platelet decrease or neutropenia is single-digit % and quite not frequent, and we have proven that it's not dose-dependent. This safety profile is really a best-in-class safety profile and allows us to combine with a multi-agent cytotoxic chemotherapy, for example, in combination with HERCEPTIN with chemotherapy, when combined with KEYTRUDA with chemotherapy. In terms of efficacy, at this moment, we have five independent combination studies, the five of them with results above the benchmark. The first one is for NHL patients in combination rituximab, patients that have had rituximab already.
We don't have a very good benchmark for this indication. Here, it stays at probably in the range of 10%, but nobody has done a randomized study of rituximab in the second-line or later after rituximab. We can compare ourselves with obinutuzumab, the similar patient population. The best response rate obinutuzumab had was 48%, and we had a 70% response rate at our higher dose of 15 mg/ kg weekly. In second-line head and neck for patients that are checkpoint-naïve, not treated with KEYTRUDA in the first-line, we had a 40% response rate. We have a benchmark for KEYNOTE-048, the phase III from Merck, benchmark is 15%, and we also do double Progression-free survival and Overall survival. In first-line head and neck, now that KEYTRUDA is approved in head and neck, we have combined with KEYTRUDA plus chemotherapy.
There, the benchmark is 36%. With a very small number of patients, only four, we had 75% at ASCO last year. We're going to update this study by the end of this year with 13 or 14 patients total. In second-line gastric cancer, patients that have failed trastuzumab in the first-line, we know that trastuzumab in the second-line does not add anything. That has been shown in a prospective phase II study that platinum plus trastuzumab had the same response rate and OS as platinum alone. In the second-line, ALX148 plus trastuzumab has a 21%, which is very significant. We put this active doublet in what is standard of care, ramucirumab, platinum. We had at OS last year a 64% response rate, where the benchmark is at 28%. We're going to update this study July 3rd at European Society for Medical Oncology Gastrointestinal.
Now we're going to go to a little more detail to each one of the studies. The first one is ALX. One point here about rituximab in combination with rituximab in NHL. In this case, rituximab binds CD20 in cancer cells and by Fc gamma receptor through macrophages, providing the required positive signal. ALX148 releases the brakes by the CD47. In slide number 17, you can see that at 10 mg/ kg, we had a 41% response rate, and at 15 mg/ kg weekly, we had a 70% response rate. This increase of response rate is statistically significant. What's interesting is that both doses will have full receptor occupancy in the periphery. They tell us that to completely block CD47 in the place of mother in the tumor, we need to reach more than 100% receptor occupancy in the periphery.
We have to have excess of antibody above the 100% receptor occupancy in blood. You can see in slide number 18, the spider plots and the waterfall plots for this trial. You can see they have responses both in indolent and aggressive lymphoma, we have partial responses and complete responses and very long duration of response. This is statistically significantly better at 15 mg/ kg compared with 10 mg/ kg. The next study in slide number 19 is gastric cancer second-line, patients that have failed trastuzumab in their first-line. In this case, trastuzumab binds HER2 on cancer cells and provides a positive signal, and ALX148 again releases the brakes. The first study we did was ALX plus trastuzumab alone in the second-line, where we know that trastuzumab by itself does not add anything, and we have a 21% response rate.
From there, we went to the second-line on top of the standard of care in slide 21, and you can see that we have associated last year a 64% response rate, which is very significant when the benchmark is 28% from the RAINBOW study. We're going to update this study by the end of July at ESMO GI. Beginning of July. From that, we're moving into a randomized phase II study in combination with trastuzumab, X eloda, platinum. There could be potential for this situation. Another study that we have done is in head and neck in combination with KEYTRUDA. Here, the mechanism is slightly different in slide number 23. SIRPalpha is now expressed in dendritic cells and CD47 in T cells. By blocking CD47, we activate dendritic cells.
Dendritic cells can present antigen to T cells, and then T cells get activated, and when we unleash them with their own checkpoint, with a PD-1 blocker, these T cells can kill cancer cells. Basically, in this case, we are helping the tumor microenvironment to create or activate T cells. We do have a very good benchmark for this study in slide 24. We know in the second-line, patients that's checkpoint-naïve, there's a 15% response rate. In the first-line, with KEYTRUDA alone is 17%, and KEYTRUDA plus chemotherapy, it is 36%. Now, our first study that was second-line before KEYTRUDA was approved in the first-line, we had 10 patients that were checkpoint-naïve and 10 patients that checkpoint-experienced. If we focus on the patients that are checkpoint-naïve in blue, you can see we have four out of 10 that responded. It's a 40% response rate.
That compares very well to the expected 15% in this line. As we double the PFS and OS. You can see we have responses in patients with very low PD-1 scores. For example, the fifth patient from the right is a combined positive score of zero. From that, we went to first line, now that pembrolizumab is approved in the first line on top of cetuximab chemotherapy. At [audio distortion] last year, we only had four patients. Out of those four, three responded, one CR and two PRs, and one of them was a CPS score of zero. Again, very promising data, and we're going to update this study by the end of the year with more patients, something like 13, 14 patients.
We are moving into two phase II studies in the first line, one in combination with KEYTRUDA and one in combination with KEYTRUDA chemotherapy. Both are randomized, both are substrate for potentially accelerated approval. We have already dosed the first patient in the top study in the ALX-KEYTRUDA. Shortly, we expect to dose the bottom one, ALX-KEYTRUDA plus chemotherapy. Finally, another mechanism that we're exploring is in combination with azacitidine and MDS. In this case, azacitidine upregulates a molecule called calreticulin in cancer cells. Calreticulin binds a receptor on macrophages called SIRPα and provides the required positive signal. Again, blocking ALX148 unleashes macrophages to eat MDS cells. Magrolimab, the kindred molecule, had very significant responses in this indication, in this combination.
You can see on the table on the left that magrolimab, in combination with azacitidine, had a 42% CR rate, when the benchmark by azacitidine alone is 17%. Very meaningful response rates for patients. As a monotherapy in the middle, magrolimab did not show much with a 0% CR. Despite the responses are really good, at the same time, magrolimab does have significant thrombocytopenia, neutropenia, and anemia. Cytopenias are important for MDS patients. Many of them will die of complications due to cytopenias. What we want to do here is make a molecule, ALX148, that will provide the same or better response rates as magrolimab, but with better safety profile. We have shown in preclinical models that azacitidine enhances calreticulin, and that ALX148 enhances activity of azacitidine in phagocytosis assays in the slide number 30.
We have shown also in preclinical models of leukemia that ALX148 can enhance activity of azacitidine in slide 31. Now, in slide 32, the clinical study that we are performing. We are now in the middle of phase I, dose escalating, to prove the safety of ALX148 in combination with azacitidine, and from that, we're going to move into a randomized phase II that could be substrate for accelerated approval. Finally, the summary for ALX148. We have shown that ALX148 has a very good safety profile that enables a combination with a wide range of agents, allows higher dosing, and a smaller molecular weight may facilitate better tumor penetration. We have clinical proof of principle, both in hematological and solid tumors. ALX148 is the only CD47 blocker to show encouraging data in solid tumors so far.
In terms of milestones for the rest of the year, as I said, we're going to disclose the data from digestive cancer phase I combination with chemotherapy at ESMO GI, and for head and neck, the phase I will be disclosed in the second half of the year. We just initiated the phase II for head and neck, and shortly we'll initiate the phase I with zanidatamab in breast, in collaboration with Zymeworks. In the second half of the year, we're going to start the phase II for gastric cancer in combination with azacitidine. We're going to start the phase I in AML, and by the end of the year, we hope to disclose data for MDS in combination with azacitidine and initiate a phase II. The results from the phase I from the AML should be coming first quarter of next year.
Now, the few minutes that we have left, have a very quick introduction to our second compound, our SIRPα TRAAC. We are really specialized in the CD47 SIRPα pathway, we understand the biology very well. SIRPα is expressed in dendritic cells, which are very important to activate T cells and have good main orchestrators of the immune system. What we have done in collaboration with another company, Tallac Therapeutics, is an anti-SIRPα antibody conjugated to a TLR9 agonist antibody. TLR9 is an agonistic molecule, a cytosine-phosphat-guanine. This molecule will bind dendritic cells, activate them directly, these dendritic cells will be able to activate the tumor microenvironment. What's interesting of this molecule is it can be dosed systemically. CpGs, the TLR agonists, have a very good clinical validation from companies like [Segmed], they have to be dosed systemically.
Here, we're dosing them systemically and could be using this molecule in multiple indications. We have shown in preclinical models in slide 36 that we can activate dendritic cells, human dendritic cells, and that we have a very strong activity in syngeneic models. For example, on the right, CT26, just two doses produce full tumor regression. In MC38, in slide 37, we have proven that we provide a long-term memory of immunogenicity, memory of efficacy. Mouse that had been treated with SIRPα TRAAC and produced complete regression, when they are challenged again with MC38, the tumor does not grow. While the naive mouse, all the mouse, the tumor grow. In slide 38, the differentiation between ALX148 and SIRPα TRAAC. ALX148 is an antagonistic molecule that is designed to maximize activity of a wide array of anticancer agents by blocking the CD47 myeloid checkpoint.
The removal of this CD47 signaling requires full blockade of the pathway. On the bottom, SIRPα TRAAC is an agonistic molecule that directly activates dendritic cells and initiates a coordinated innate and adaptive immune response against cancer. In the case of agonistic molecules, constant blockade is not required. In terms of positioning, we think about ALX148 to go on top of the standard of care. We think about the SIRPα after standard of care, when standard of care fails. Finally, in slide 39, our financial information. As of March 31st, 2021, we have $429 million in the bank, which is expected to runway through 2024, that's enough to finish all the phase II that we just described. Thank you for your attention.