WebGLRenderingContext's texImage2D and texSubImage2D question

[MOLTogCHIN]

Hello Mattdesl, I have a one question about WebGLRenderingContext's texImage2D and texSubImage2D function.
What I trying to create is something similar like loading image by tile by tile.

Here is my first function. It's working properly while it's first load. But whenever loading second time using same function but tiled background.
It failed. Look at the code below

// This is first load code which is working fine

gl.bindTexture(gl.TEXTURE_2D, handle)
gl.texImage2D( gl.TEXTURE_2D, 0, gl.RGBA, width, height, 0, gl.RGBA, gl.UNSIGNED_BYTE, null)

gl.bindTexture(gl.TEXTURE_2D, handle)
gl.texSubImage2D(gl.TEXTURE_2D, 0, x, y, gl.RGBA, gl.UNSIGNED_BYTE, image)

// This is second time using same but one different coding section [image data from canvas]. But it's not working.

gl.bindTexture(gl.TEXTURE_2D, handle)
gl.texImage2D( gl.TEXTURE_2D, 0, gl.RGBA, width, height, 0, gl.RGBA, gl.UNSIGNED_BYTE, [image data from canvas])

gl.bindTexture(gl.TEXTURE_2D, handle)
gl.texSubImage2D(gl.TEXTURE_2D, 0, x, y, gl.RGBA, gl.UNSIGNED_BYTE, image)

Have you ever face the problem similar like me. Then please guide me to the light. Much appreciate it.